SlideShare a Scribd company logo
1 of 40
Download to read offline
© 2010 IBM Corporation
Conference materials may not be reproduced in whole or in part without the prior written permission of IBM.
WebSphere Performance for MulticoreWebSphere Performance for Multicore
and Virtualised Platformsand Virtualised Platforms
Session Number: W22Session Number: W22
Hendrik van RunHendrik van Run –– hvanrun@uk.ibm.comhvanrun@uk.ibm.com
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
2
Agenda
• Multicore processor technology
Why do we need/have multicore processors?
Examples of Multicore processor architectures
How to get value out of Multicore platforms?
• Known Issues and Best Practices for Multicore
Administration
Programming
• Virtualisation in the Enterprise
Why Virtualisation?
Different forms of Virtualisation
Where to start?
• Known Issues and Best Practices for Virtualisation
Performance Overhead
Resource Over-Commitment
WebSphere Performance Best Practices
Other Best Practices
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
3
Why do we need/have multicore processors?
Moore’s law
• Moore's law describes a long-
term trend in the history of
computing hardware
The number of transistors
that can be placed
(inexpensively) on an
integrated circuit has doubled
approximately every two
years
The trend has continued for
more than half a century
This is not expected to stop
until 2015 or later
• For almost two decades, single thread performance has doubled every 2
years in a very predictable fashion
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
4
Why do we need/have multicore processors?
CPU clockspeed has hit a wall
• Performance has
been propelled
forward by
continuing
advancements in
Clock speed
Instruction Level
Parallelism (ILP)
L1/L2/L3 cache
size
• Heat dissipation increases linearly with processor frequency
This is the main reason that clock speeds cannot easily exceed 3-5 GHz
• Using mutiple cores on the same integrated circuit continues to allow for
increased performance
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
5
Why do we need/have multicore processors?
The multicore future: System on a chip dreams
• Massively complex and interesting problems await chip
designers
• How else can we use all these extra transistors?
More processors on a chip (Multi-core)
More threads on a core (SMT)
Special-purpose processors (eg. crypto, compression)
Integrate system components on the chip (eg. network, I/O)
• Multicore is only the first, most notable impact of the
physical phenomena of ever shrinking die sizes
• The real story is that the measure of success for future processors is changing
It’s no longer about being the Gigahertz champion or having the most clever microarchitecture
It’s about who can use their transistor budget most effectively to speed up real workloads
• One size will not fit all
Greater focus is required on current and emerging workload characteristics
Both general purpose and special purpose processing will play a role
Processor and system effectiveness will be measured more aggressively against multiple criteria
(performance, power, price, reliability)
Greater levels of integration in hardware will drive new requirements into software
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
6
Why do we need/have multicore processors?
What are the benefits of multicore platforms?
• Consolidation
Most admins are looking at the sheer number of cores in a single box now as a means
to consolidate older larger systems
Allows for collapsing down of multi-tiered architectures
• Power Savings
Power costs in a system come from more than just the processor.
Ability to consolidate down 4 machines into 1 will save you power.
Eliminating machines from your infrastructure provides consolidation of the sheer
amount of moving parts in your infrastructure saving you money.
• Performance
Moving from old single core hardware to new multicore hardware can result in
significant performance improvements to legacy applications
• All of the above effectively reduce cost
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
7
Why do we need/have multicore processors?
Why should I understand multicore platforms?
• Typical hardware replacements cycles mean boxes with these processors are coming into
your IT shop soon if not already there
• These systems present unique challenges for Operating System, WebSphere, and
Application Deployment outside of the past 10 years of knowledge
• Hardware specific features and characteristics can cause major performance swings in
applications
• Requires developers to understand how to write code optimized for deployments on this
type of hardware or/and Administrators that know how to work around multicore code
problems.
• Tighter coupling of operations and development is required
• Through understanding of 3rd party libraries
• Possible reworking of existing applications required to make them run optimally on
micropartitioned or over committed hardware
• Intelligent and insightful planning will ease your pain
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
8
Why do we need/have multicore processors?
Examples of Multicore processor architectures
Intel Nehalem
• 1.86 – 3.6 GHz
• 2 – 8 cores per chip
• Up to 2 threads per
core (SMT)
IBM Power7
• 3.0 – 4.25 GHz
• 4 – 8 cores per chip
• Up to 4 threads per
core (SMT)
IBM Power6
• 3.5 – 5.0 GHz
• 2 cores per chip
• Up to 2 threads per
core (SMT)
AMD Phenom II
• 2.5 – 3.5 GHz
• 2 – 6 cores per chip
• Up to 2 threads per
core (SMT)
Sun Niagara 2
• 1.2 – 1.6 GHz
• 4 – 8 cores per chip
• Up to 8 threads per
core (SMT)
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
9
How to get value out of Multicore platforms?
Virtualisation
• Large number of cores can be a challenge
How can you make sure that you’re using them all?
• Virtualisation enables you to effectively run more applications on the
same machine
Different types of virtualisation exist
We will discuss those in more detail towards the end of this presentation
• There is no single “best” virtualisation technology out there
Depends very much on your business and application needs
• Start with requirements
Do you want movable OS images?
Do you need OS level partitioning?
Do you need Dynamic resource allocation?
Do you want to simply pay the lowest cost?
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
10
How to get value out of Multicore platforms?
Understand scaling of WebSphere Application Server
• The evolution of multicore hardware places unique demands
on application server performance
A single instance must perform well at a low number of cores,
but also scale to take advantage of many cores
Multiple instances must be able to run concurrently providing
“vertical scaling” on a single server
• WAS V7 Multicore Performance
WAS V7 achieves close to 1000 requests/second on a single
CPU core running the DayTrader benchmark in this configuration
WAS is able to take advantage of all 12 cores showing significant
performance increases as more cores are enabled
The ability to achieve high scalability depends on application
architecture
• WAS V7 Multicore Scalability
The 2nd graph depicts WAS V7 scalability relative to a linear
scale.
With this benchmark, WAS V7 scales well to 8 cores. Scalability
between 8 and 12 cores is beginning to flatten, but significant
improvement is still being measured at 12 cores
• WAS V7 Vertical Scaling
WAS V7 ND provides the foundation and features for developing
a vertically clustered application (or applications) on a single
platform
Vertical scaling can take advantage of many CPU cores and
dramatically improve single server scalability
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
11
Known Issues and
Best Practices for Multicore
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
12
Known Issues and Best Practices for Multicore
Introduction
• Multicore platforms require specific knowledge
Administrators
Developers
Architects
• Single thread application performance is typically lower
Most Multicore processors run at a relative low clockspeed
This fundamental limitation is central to what we will discuss here
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
13
Known Issues and Best Practices for Multicore
Administration
• Multicore processor designs focus on building more capacity
No longer an increase in clockspeed with every new processor
• Clock speeds are up against the boundary of what’s possible
Effects anything that is single threaded:
Application Startup Time (i.e. 60 s instead of 15 s)
WSAdmin and scripting runtime startup
Batch compilation of JSPs
Client Response Times (i.e. 1.5 s instead of 500 ms)
• Hence you can no longer expect performance to double every 2 years
Incremental improvements in processor design will be made optimizing
pipelines and instruction cycles
Unlikely to see similar increase from the past in single thread performance
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
14
Known Issues and Best Practices for Multicore
Administration
• Memory capacity growth is slower than CPU capacity growth
Most of the new multicore systems are limited in memory
Forced to run 8-16 hardware threads or more against 8GB-16GB of
space
Be very careful not to overcommit your memory resources!
5% CPU utilisation and 100% memory consumption is common
• Very advanced and sensitive BIOS and other hardware
settings required to get the most performance
SMT/CMT/HT effect application performance
Intel/AMD “Turbo Mode” has drastic effects on different applications
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
15
Known Issues and Best Practices for Multicore
Administration
• Consolidated environments can lead to issues with High
Availability
Less machines means a single outage results in more applications
effected
Need to plan to ensure data replication occurs off a single machine
• Legacy software can suffer massively
Massive concurrent load on code not designed or tested through
enough to handle it
Clock speed drop effects CPU intensive code paths
• Hardware while becoming much more powerful is becoming
more specialized
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
16
Known Issues and Best Practices for Multicore
Administration
• WAS administrative best practices
With previously discussed memory and CPU speed concerns
Ensure your heaps and thread pools are correctly sized
With WSAdmin scripts limit the initialization of the framework
Ensure performance acceptance testing is completed on identical
deployment topology
Business level application planning should take into consideration
hardware variations
Work with developers to institute codepath focused profiling and
improvement in applications
• Garbage Collection Overhead
Due to the increase in processing speed and capability on a single
system the amount of garbage created in a single JVM increases and
you must tune appropriately
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
17
Known Issues and Best Practices for Multicore
Programming
• Programming for a multicore world is not THAT much different than
writing good performing code today.
Developers just need to do a little more due diligence before committing code
Develop with the deployment hardware in the back of your head. You cannot
treat the hardware as an unknown
• Maximize single thread (request) performance
Minimizes response time for end user requests
Code for a request that can execute as efficiently as possible on a single
thread leads to better performance and in most cases better scalability as
concurrent requests go up
Developers have to be more aware of what their code is doing as it flows
through 3rd party libraries
What to do:
Focus on requests traversal from the socket down to the backend system
Review your code to remove unnecessary caching or CPU intensive method calls
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
18
Known Issues and Best Practices for Multicore
Programming
• Utilize Effective Memory Management
The JVM for all of its advancements still has its performance dictated to it by
garbage collection
Minimize Object creation by avoiding the overly object oriented scenario as well
as minimizing reflection. Keep it primitive where ever possible.
Audit your code for calls to “new” but avoid pooling except for excessively
complex object trees.
If exploiting internal pooling or caching ensure it happens as high up the code
stack as possible.
Keep object structures as “thin” as possible.
• Exploit Advanced Synchronization
To be able to scale up effectively with the move to multicore hardware every
developer needs to be well aware of explicit and implicit synchronization
Define early on what data need to be shared across hardware threads and what
doesn’t and use appropriate data structures.
Ensure synchronization is at the lowest possible level. Decompose locking
schemes.
For shared data structures utilize java.util.conncurrent or advanced libraries
Expose outwardly via java docs the level of synchronization to API level calls to
your component for internal consumption.
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
19
Known Issues and Best Practices for Multicore
Programming
• Exploit Effective Use of Debugging, Trace, and Logging Code
Minimizes runtime codepath length and GC overhead
The JIT is excellent at optimizing code out but it is still prudent to create
clean code it is also critical to create performant trace code
What to do:
Create trace code for your method entry and exits but ensure they are properly
gated
Review your trace code to make sure it is putting out information to help with
debugging but nothing more to keep overhead as low as possible
• Remove unused or redundant code
Reduces startup time by cutting the amount of class size being loaded from
disk
Redundant code and unused code contributes to class file size growth
resulting in classload times increases
What to do:
Audit your code to ensure that all the method calls in your classes are actually
accessed from other code, if not remove them.
Ensure that in your methods all code is reachable by running one of the various
Eclipse plugins that detects dead and unreachable code.
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
20
Virtualisation in the Enterprise
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
21
Why Virtualisation?
• Most datacenters are heavily underutilized with systems running
between 5-10% CPU on average
This is not intentional but typically due to the unfounded need for all
applications to run on discrete machines OR applications cyclical life
cycles
All modern hardware can easily run at above 50% CPU without
losing performance
• Not running your servers at maximum capacity is costing you
dollars in terms of software licenses, power consumption, floor
space, etc
• There is no single solitary way to virtualize your environment. You
need a strategy based off of your applications and business needs
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
22
Why Virtualisation?
• Consolidation
Most admins are looking at the sheer number of cores in a single box now as a means
to consolidate older larger systems
Allows for collapsing down of multi-tiered architectures
• Power Savings
Power costs in a system come from more than just the processor.
Ability to consolidate down 4 machines into 1 will save you power.
Eliminating machines from your infrastructure provides consolidation of the sheer
amount of moving parts in your infrastructure saving you money.
• Performance
Moving from old single core hardware to new multicore hardware can result in
significant performance improvements to legacy applications
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
23
Why Virtualisation?
• Typical hardware replacements cycles mean boxes with
these processors are coming into your IT shop soon if not
already there so take advantage of them!!
• Virtualization can drastically cut down on the energy and
effort to get new WAS environments stood up for testing and
production
• In the world of every tighter budgets and constrained
resources the ability to get the most out of your hardware
makes a difference on the bottom line.
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
24
Different forms of Virtualisation
Hardware Virtualization
• Advantages
– Pure electronic
isolation
– Hardware based
resource sharing
– Higher Performance
• Disadvantages
– Management and
maintenance of large
topology of systems
Software Virtualization OS Virtualization
Power7 Hardware
• Advantages
– Hardware agnostic
moveable OS images
– OS based isolation
– Software based
resource sharing
– Good Performance
• Disadvantages
– I/O system
performance
– System Tooling not
accurate
• Advantages
– Single OS installation
– OS isolation/security
– Simplistic box based
administration
– Near on metal
performance
• Disadvantages
– System level only
management
– Dependent on base
LPAR/Hardware for
availability
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
25
Different forms of Virtualisation
Multiple WebSphere
Instances
Single WAS multiple
applications
WAS
App
App
App
App
App
App
App
App
• Advantages
– Single OS installation
– Process based security
– Fast install, setup and boot
– Best performing setup
– Least memory consumption
• Disadvantages
– Process level security only
• Advantages
– Single OS installation
– Simplistic Administration
– Good performance
• Disadvantages
– No application isolation
– Rouge Applications
– Minimal availability
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
26
Where to start with Virtualisation?
• Start small and work big towards a grander plan
• By starting small and slowly rolling applications from “bare metal” into your virtualized
infrastructure you can work through issues and adjust future plans
• By having a grand vision on where the datacenter in your business is heading its
easier to sell the long term value and show short term progress in real numbers
• Begin by working with back office, old non critical, new applications, or development
• New applications
No known baseline to be compared to so no one can say “the old stuff was faster”
Can be build and tested from the beginning with virtualization in mind. Thus all performance
testing, etc can be done on like systems.
• Back office / Old non critical applications
Nice starting point as huge amount of information on access patterns and other factors
affecting performance exist.
Minimal risk of impacting large user sets
• Gain knowledge and roll that knowledge into your long term plan
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
27
Known Issues and
Best Practices for Virtualisation
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
28
Known Issues and Best Practices for Virtualisation
Virtualisation Performance Overhead
• Performance overhead differs by platform and virtualization technology
• Completely virtualized environments typically have higher performance penalties
Network and disk IO latencies significantly higher in these environments
• Para-virtualization technologies provide an optimized path to the underlying
hardware for these high latency paths
VMware Tools provide this capability in VMware environments
Optimizations for virtual Ethernet driver, memory sharing and time sync
Enabling VMI kernel support on VMware/Linux-based guests reduces virtualization
overhead further
Check VMware documentation for supported kernels
• Newer hardware technologies provide better virtualization support and
performance
For example: Intel VT and AMD-V
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
29
Known Issues and Best Practices for Virtualisation
Virtualisation Performance Overhead – VMware ESX 3.5
• Example overhead measurements based on DayTrader application
Minimal disk I/O characteristics
Increased I/O activity will increase performance delta
• VMI kernel reduces overhead further (~ 5% overhead)
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
30
Known Issues and Best Practices for Virtualisation
Virtualisation Performance Overhead – PowerVM LPARs
• Traditional “Thick” LPARs provide hard isolation and dedicated resources
“Thin” LPARs use the AIX Network Installation Manager (NIM) to share a single
remote OS boot image
• Minimal disk I/O characteristics are behind the small performance delta
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
31
Known Issues and Best Practices for Virtualisation
Virtualisation Performance Overhead – PowerVM WPARs
• Workload Partitions (WPARs) provide a virtualized operating environment activated within the
global OS
• Virtualization overhead minimal
• Multi-instance scalability impacted by normal process scheduling hurdles
RSET can be used to achieve optimal throughput for multiple instances in both LPAR and WPAR
environments
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
32
Known Issues and Best Practices for Virtualisation
Virtualisation Performance Overhead – Different Levels of Isolation
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
33
Known Issues and Best Practices for Virtualisation
Resource Over-Commitment
• Memory
Never over commit physical memory in the system if you expect predictable performance
When pushing the boundaries of physical RAM in a system smaller JVM heaps allow for more WAS
instances on a single hardware machine
• CPU
CPU over commit results in good performance up to the point of 100% saturation on the hardware
At which point, response time begin to suffer
Limit Memory Over commit Lower CPU Utilization Over commit
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
34
Known Issues and Best Practices for Virtualisation
WebSphere Performance Best Practices
• Thread pool settings
Must be sized not to the number of virtual CPUs exposed but to the
number of physical CPUs on the system
• Heap Sizes
Set heap sizes to ensure that guest remain in physical memory for
optimal performance.
Right size your heap to ensure you are not wasting memory
• High Availability
You must implement a intelligent placement mechanism to ensure
fault tolerance for your applications
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
35
Known Issues and Best Practices for Virtualisation
WebSphere Performance Best Practices
• Deployment Scripting
Script your administrative tasks to ensure that installation and
configuration of multiple nodes is repeatable
Allows you to bundle and distribute images quickly and have them up
and running in a “certified” manner
• Minimizing IO
Minimize the application server IO to ensure maximum performance
Cut logging to a bare minimum
Bare minimum writing to disk should be done unless the disk system
is optimized
Ensure “chattiness” of your application is as minimal as possible
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
36
Known Issues and Best Practices for Virtualisation
Other Best Practices
• Resource utilization monitoring
Virtualization often skews the CPU, IO and other resources metrics
reported by the Guest OS
Need to rely on the Hypervisor for resource utilization statics
Example:
AIX micro-partitioning
• Clock sync / drift
Problem noticed in VMware / WebSphere HA environments
Installation of VMware Tools and use of VMI kernel address this
issue
• VM migration tools (ie. VMotion)
Avoid these technologies in your environment
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
37
Known Issues and Best Practices for Virtualisation
Other Best Practices
• WebSphere tuning can be directly linked to resources allocated to Guest
VM
• Reduce disk space and memory requirements for Guests by disabling or
removing un-used services and applications
• Leverage VM cloning and templating technologies to provide a uniform
operating environment
For example: “Linked-clone Templates” in VMware
Scripts for changing WAS hostnames
http://www.ibm.com/developerworks/websphere/techjournal/0905_webcon/0
905_webcon.html
• Use SANs or other fast disk technologies for IO intensive workloads
• Can I trust Java profiling tools in virtualized environments?
Successfully used Java Health Center CPU profiling to diagnose
performance issue
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
38
Deploy Applications in Virtualised Environments
WebSphere CloudBurst 2.0 Appliance
• Ability to rapidly deploy IBM middleware and customer applications into virtual
and cloud environments leveraging best practice deployment patterns
• Support for multiple hypervisor platforms, such as VMware ESX, PowerVM,
and z/VM
38
• Availability of the following product images
WebSphere Application Server Hypervisor Edition
V6.1 and V7
WebSphere Application Server Hypervisor Edition V7
now available on Red Hat Enterprise Linux!
WebSphere Application Server Hypervisor Edition -
Intelligent Management Pack
Available for WebSphere Application Server Hypervisor
Edition V7
WebSphere Process Server Hypervisor Edition V6.2
and V7
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
39
Summary
• Multicore processor technology and Virtualisation
Technologies are here today
And this trend will continue…
• WebSphere performs well on those platforms
But make sure to address known issues and apply best practices
WebSphere Technical Conference and Portal
Excellence Conference
© 2010 IBM Corporation
40
QuestionsQuestions
• Please remember to complete your session evaluation form
Please complete your session evaluation, session number is W22

More Related Content

What's hot

1.ibm pure flex system mar 2013
1.ibm pure flex system   mar 20131.ibm pure flex system   mar 2013
1.ibm pure flex system mar 2013solarisyougood
 
Linux and ibm client ready strategy pitch - april 2012
Linux and ibm client ready strategy pitch - april 2012Linux and ibm client ready strategy pitch - april 2012
Linux and ibm client ready strategy pitch - april 2012systemxteambelux
 
AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...
AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...
AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...WASdev Community
 
Flash Ahead: IBM Flash System Selling Point
Flash Ahead: IBM Flash System Selling PointFlash Ahead: IBM Flash System Selling Point
Flash Ahead: IBM Flash System Selling PointCTI Group
 
A Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and BeyondA Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and BeyondLaura Hood
 
Flex system client_presentation
Flex system client_presentationFlex system client_presentation
Flex system client_presentationNatalija Pavic
 
Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013
Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013
Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013Amazon Web Services
 
Mainframe cloud computing presentation
Mainframe cloud computing presentationMainframe cloud computing presentation
Mainframe cloud computing presentationxKinAnx
 
IBM flash systems
IBM flash systems IBM flash systems
IBM flash systems Solv AS
 
Key Note Session IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...
Key Note Session  IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...Key Note Session  IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...
Key Note Session IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...Surekha Parekh
 
VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...
VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...
VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...VMworld
 
Ibm symp14 referent_philipp kessler_storage virtualisierung
Ibm symp14 referent_philipp kessler_storage virtualisierungIbm symp14 referent_philipp kessler_storage virtualisierung
Ibm symp14 referent_philipp kessler_storage virtualisierungIBM Switzerland
 
Planning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPMPlanning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPMWASdev Community
 
VMworld 2016: What's New with Horizon 7
VMworld 2016: What's New with Horizon 7VMworld 2016: What's New with Horizon 7
VMworld 2016: What's New with Horizon 7VMworld
 
Speed up the cloud adoption with SoftLayer Cloud Services - dominopoint
Speed up the cloud adoption with SoftLayer Cloud Services - dominopointSpeed up the cloud adoption with SoftLayer Cloud Services - dominopoint
Speed up the cloud adoption with SoftLayer Cloud Services - dominopointDominopoint - Italian Lotus User Group
 
SemeruRuntimesUnderTheCover .pptx
SemeruRuntimesUnderTheCover .pptxSemeruRuntimesUnderTheCover .pptx
SemeruRuntimesUnderTheCover .pptxSumanMitra22
 
IBM Endpoint Manager for Server Automation presentation
IBM Endpoint Manager for Server Automation presentationIBM Endpoint Manager for Server Automation presentation
IBM Endpoint Manager for Server Automation presentationRMayo22
 
Horizon view technical deep dive
Horizon view   technical deep diveHorizon view   technical deep dive
Horizon view technical deep diveMurugesan Arumugam
 

What's hot (20)

Designing virtual infrastructure
Designing virtual infrastructureDesigning virtual infrastructure
Designing virtual infrastructure
 
1.ibm pure flex system mar 2013
1.ibm pure flex system   mar 20131.ibm pure flex system   mar 2013
1.ibm pure flex system mar 2013
 
Patterns
Patterns Patterns
Patterns
 
Linux and ibm client ready strategy pitch - april 2012
Linux and ibm client ready strategy pitch - april 2012Linux and ibm client ready strategy pitch - april 2012
Linux and ibm client ready strategy pitch - april 2012
 
AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...
AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...
AAI-3281 Smarter Production with WebSphere Application Server ND Intelligent ...
 
Flash Ahead: IBM Flash System Selling Point
Flash Ahead: IBM Flash System Selling PointFlash Ahead: IBM Flash System Selling Point
Flash Ahead: IBM Flash System Selling Point
 
A Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and BeyondA Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
A Time Traveller's Guide to DB2: Technology Themes for 2014 and Beyond
 
Flex system client_presentation
Flex system client_presentationFlex system client_presentation
Flex system client_presentation
 
Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013
Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013
Network-Ready Your Hybrid IT Environment (ENT108) | AWS re:Invent 2013
 
Mainframe cloud computing presentation
Mainframe cloud computing presentationMainframe cloud computing presentation
Mainframe cloud computing presentation
 
IBM flash systems
IBM flash systems IBM flash systems
IBM flash systems
 
Key Note Session IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...
Key Note Session  IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...Key Note Session  IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...
Key Note Session IDUG DB2 Seminar, 16th April London - Julian Stuhler .Trito...
 
VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...
VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...
VMworld Europe 2014: Ask the Experts - Design Advice for Small and Midsize Bu...
 
Ibm symp14 referent_philipp kessler_storage virtualisierung
Ibm symp14 referent_philipp kessler_storage virtualisierungIbm symp14 referent_philipp kessler_storage virtualisierung
Ibm symp14 referent_philipp kessler_storage virtualisierung
 
Planning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPMPlanning For Catastrophe with IBM WAS and IBM BPM
Planning For Catastrophe with IBM WAS and IBM BPM
 
VMworld 2016: What's New with Horizon 7
VMworld 2016: What's New with Horizon 7VMworld 2016: What's New with Horizon 7
VMworld 2016: What's New with Horizon 7
 
Speed up the cloud adoption with SoftLayer Cloud Services - dominopoint
Speed up the cloud adoption with SoftLayer Cloud Services - dominopointSpeed up the cloud adoption with SoftLayer Cloud Services - dominopoint
Speed up the cloud adoption with SoftLayer Cloud Services - dominopoint
 
SemeruRuntimesUnderTheCover .pptx
SemeruRuntimesUnderTheCover .pptxSemeruRuntimesUnderTheCover .pptx
SemeruRuntimesUnderTheCover .pptx
 
IBM Endpoint Manager for Server Automation presentation
IBM Endpoint Manager for Server Automation presentationIBM Endpoint Manager for Server Automation presentation
IBM Endpoint Manager for Server Automation presentation
 
Horizon view technical deep dive
Horizon view   technical deep diveHorizon view   technical deep dive
Horizon view technical deep dive
 

Similar to W22 - WebSphere Performance for Multicore and Virtualised Platforms

Multicore 101: Migrating Embedded Apps to Multicore with Linux
Multicore 101: Migrating Embedded Apps to Multicore with LinuxMulticore 101: Migrating Embedded Apps to Multicore with Linux
Multicore 101: Migrating Embedded Apps to Multicore with LinuxBrad Dixon
 
Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...
Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...
Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...Andrew Miller
 
Virtualization and Cloud Computing
Virtualization and Cloud ComputingVirtualization and Cloud Computing
Virtualization and Cloud ComputingJosh Folgado
 
Presentation architecting a cloud infrastructure
Presentation   architecting a cloud infrastructurePresentation   architecting a cloud infrastructure
Presentation architecting a cloud infrastructurexKinAnx
 
Presentation architecting a cloud infrastructure
Presentation   architecting a cloud infrastructurePresentation   architecting a cloud infrastructure
Presentation architecting a cloud infrastructuresolarisyourep
 
Azure enterprise integration platform
Azure enterprise integration platformAzure enterprise integration platform
Azure enterprise integration platformMichael Stephenson
 
SCU 2015 - My top 10 favorite items you need to look at in WK2012R2
SCU 2015 - My top 10 favorite items you need to look at in WK2012R2SCU 2015 - My top 10 favorite items you need to look at in WK2012R2
SCU 2015 - My top 10 favorite items you need to look at in WK2012R2Mike Resseler
 
A Software Factory Integrating Rational & WebSphere Tools
A Software Factory Integrating Rational & WebSphere ToolsA Software Factory Integrating Rational & WebSphere Tools
A Software Factory Integrating Rational & WebSphere Toolsghodgkinson
 
GTC15-Manoj-Roge-OpenPOWER
GTC15-Manoj-Roge-OpenPOWERGTC15-Manoj-Roge-OpenPOWER
GTC15-Manoj-Roge-OpenPOWERAchronix
 
Ibm PureApplication system
Ibm PureApplication systemIbm PureApplication system
Ibm PureApplication systemkhawkwf
 
29092013042656 multicore-processor-technology
29092013042656 multicore-processor-technology29092013042656 multicore-processor-technology
29092013042656 multicore-processor-technologySindhu Nathan
 
Virtual Design Master Challenge 1 - Akmal
Virtual Design Master Challenge 1  - AkmalVirtual Design Master Challenge 1  - Akmal
Virtual Design Master Challenge 1 - Akmaltovmug
 
Designing your xen app 7.5 environment
Designing your xen app 7.5 environmentDesigning your xen app 7.5 environment
Designing your xen app 7.5 environmentsolarisyougood
 
Stay productive while slicing up the monolith
Stay productive while slicing up the monolithStay productive while slicing up the monolith
Stay productive while slicing up the monolithMarkus Eisele
 
Software Factories in the Real World: How an IBM WebSphere Integration Factor...
Software Factories in the Real World: How an IBM WebSphere Integration Factor...Software Factories in the Real World: How an IBM WebSphere Integration Factor...
Software Factories in the Real World: How an IBM WebSphere Integration Factor...ghodgkinson
 
Stay productive while slicing up the monolith
Stay productive while slicing up the monolithStay productive while slicing up the monolith
Stay productive while slicing up the monolithMarkus Eisele
 
ODSA Sub-Project Launch
ODSA Sub-Project LaunchODSA Sub-Project Launch
ODSA Sub-Project LaunchODSA Workgroup
 
ODSA Sub-Project Launch
 ODSA Sub-Project Launch ODSA Sub-Project Launch
ODSA Sub-Project LaunchNetronome
 

Similar to W22 - WebSphere Performance for Multicore and Virtualised Platforms (20)

Multicore 101: Migrating Embedded Apps to Multicore with Linux
Multicore 101: Migrating Embedded Apps to Multicore with LinuxMulticore 101: Migrating Embedded Apps to Multicore with Linux
Multicore 101: Migrating Embedded Apps to Multicore with Linux
 
Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...
Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...
Varrow Q4 Lunch & Learn Presentation - Virtualizing Business Critical Applica...
 
Virtualization and Cloud Computing
Virtualization and Cloud ComputingVirtualization and Cloud Computing
Virtualization and Cloud Computing
 
TechTalkThai-CiscoHyperFlex
TechTalkThai-CiscoHyperFlexTechTalkThai-CiscoHyperFlex
TechTalkThai-CiscoHyperFlex
 
Deco1
Deco1Deco1
Deco1
 
Presentation architecting a cloud infrastructure
Presentation   architecting a cloud infrastructurePresentation   architecting a cloud infrastructure
Presentation architecting a cloud infrastructure
 
Presentation architecting a cloud infrastructure
Presentation   architecting a cloud infrastructurePresentation   architecting a cloud infrastructure
Presentation architecting a cloud infrastructure
 
Azure enterprise integration platform
Azure enterprise integration platformAzure enterprise integration platform
Azure enterprise integration platform
 
SCU 2015 - My top 10 favorite items you need to look at in WK2012R2
SCU 2015 - My top 10 favorite items you need to look at in WK2012R2SCU 2015 - My top 10 favorite items you need to look at in WK2012R2
SCU 2015 - My top 10 favorite items you need to look at in WK2012R2
 
A Software Factory Integrating Rational & WebSphere Tools
A Software Factory Integrating Rational & WebSphere ToolsA Software Factory Integrating Rational & WebSphere Tools
A Software Factory Integrating Rational & WebSphere Tools
 
GTC15-Manoj-Roge-OpenPOWER
GTC15-Manoj-Roge-OpenPOWERGTC15-Manoj-Roge-OpenPOWER
GTC15-Manoj-Roge-OpenPOWER
 
Ibm PureApplication system
Ibm PureApplication systemIbm PureApplication system
Ibm PureApplication system
 
29092013042656 multicore-processor-technology
29092013042656 multicore-processor-technology29092013042656 multicore-processor-technology
29092013042656 multicore-processor-technology
 
Virtual Design Master Challenge 1 - Akmal
Virtual Design Master Challenge 1  - AkmalVirtual Design Master Challenge 1  - Akmal
Virtual Design Master Challenge 1 - Akmal
 
Designing your xen app 7.5 environment
Designing your xen app 7.5 environmentDesigning your xen app 7.5 environment
Designing your xen app 7.5 environment
 
Stay productive while slicing up the monolith
Stay productive while slicing up the monolithStay productive while slicing up the monolith
Stay productive while slicing up the monolith
 
Software Factories in the Real World: How an IBM WebSphere Integration Factor...
Software Factories in the Real World: How an IBM WebSphere Integration Factor...Software Factories in the Real World: How an IBM WebSphere Integration Factor...
Software Factories in the Real World: How an IBM WebSphere Integration Factor...
 
Stay productive while slicing up the monolith
Stay productive while slicing up the monolithStay productive while slicing up the monolith
Stay productive while slicing up the monolith
 
ODSA Sub-Project Launch
ODSA Sub-Project LaunchODSA Sub-Project Launch
ODSA Sub-Project Launch
 
ODSA Sub-Project Launch
 ODSA Sub-Project Launch ODSA Sub-Project Launch
ODSA Sub-Project Launch
 

More from Hendrik van Run

Open shift deployment review getting ready for day 2 operations
Open shift deployment review   getting ready for day 2 operationsOpen shift deployment review   getting ready for day 2 operations
Open shift deployment review getting ready for day 2 operationsHendrik van Run
 
WSI35 - WebSphere Extreme Scale Customer Scenarios and Use Cases
WSI35 - WebSphere Extreme Scale Customer Scenarios and Use CasesWSI35 - WebSphere Extreme Scale Customer Scenarios and Use Cases
WSI35 - WebSphere Extreme Scale Customer Scenarios and Use CasesHendrik van Run
 
WSI33 - Advanced Performance Tactics for IBM WebSphere Application Server
WSI33 - Advanced Performance Tactics for IBM WebSphere Application ServerWSI33 - Advanced Performance Tactics for IBM WebSphere Application Server
WSI33 - Advanced Performance Tactics for IBM WebSphere Application ServerHendrik van Run
 
WSI32 - IBM WebSphere Performance Fundamentals
WSI32 - IBM WebSphere Performance FundamentalsWSI32 - IBM WebSphere Performance Fundamentals
WSI32 - IBM WebSphere Performance FundamentalsHendrik van Run
 
W23 - Advanced Performance Tactics for WebSphere Performance
W23 - Advanced Performance Tactics for WebSphere PerformanceW23 - Advanced Performance Tactics for WebSphere Performance
W23 - Advanced Performance Tactics for WebSphere PerformanceHendrik van Run
 
2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...
2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...
2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...Hendrik van Run
 
2596 - Integrating PureApplication System Into Your Network
2596 - Integrating PureApplication System Into Your Network2596 - Integrating PureApplication System Into Your Network
2596 - Integrating PureApplication System Into Your NetworkHendrik van Run
 
1457 - Reviewing Experiences from the PureExperience Program
1457 - Reviewing Experiences from the PureExperience Program1457 - Reviewing Experiences from the PureExperience Program
1457 - Reviewing Experiences from the PureExperience ProgramHendrik van Run
 
CSD-2881 - Achieving System Production Readiness for IBM PureApplication System
CSD-2881 - Achieving System Production Readiness for IBM PureApplication SystemCSD-2881 - Achieving System Production Readiness for IBM PureApplication System
CSD-2881 - Achieving System Production Readiness for IBM PureApplication SystemHendrik van Run
 
ACU-1445 - Bringing workloads into production on PureApplication System
ACU-1445 - Bringing workloads into production on PureApplication SystemACU-1445 - Bringing workloads into production on PureApplication System
ACU-1445 - Bringing workloads into production on PureApplication SystemHendrik van Run
 
CIT-2697 - Customer Success Stories with IBM PureApplication System
CIT-2697 - Customer Success Stories with IBM PureApplication SystemCIT-2697 - Customer Success Stories with IBM PureApplication System
CIT-2697 - Customer Success Stories with IBM PureApplication SystemHendrik van Run
 
CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...
CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...
CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...Hendrik van Run
 
IC6284A - The Art of Choosing the Best Cloud Solution
IC6284A - The Art of Choosing the Best Cloud SolutionIC6284A - The Art of Choosing the Best Cloud Solution
IC6284A - The Art of Choosing the Best Cloud SolutionHendrik van Run
 
C219 - Docker and PureApplication Patterns: Better Together
C219 - Docker and PureApplication Patterns: Better TogetherC219 - Docker and PureApplication Patterns: Better Together
C219 - Docker and PureApplication Patterns: Better TogetherHendrik van Run
 
PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...
PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...
PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...Hendrik van Run
 
C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...
C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...
C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...Hendrik van Run
 
7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local System
7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local System7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local System
7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local SystemHendrik van Run
 
IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...
IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...
IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...Hendrik van Run
 
IBM Think 2019 session 2116 - Best practices for operating and managing a pro...
IBM Think 2019 session 2116 - Best practices for operating and managing a pro...IBM Think 2019 session 2116 - Best practices for operating and managing a pro...
IBM Think 2019 session 2116 - Best practices for operating and managing a pro...Hendrik van Run
 

More from Hendrik van Run (19)

Open shift deployment review getting ready for day 2 operations
Open shift deployment review   getting ready for day 2 operationsOpen shift deployment review   getting ready for day 2 operations
Open shift deployment review getting ready for day 2 operations
 
WSI35 - WebSphere Extreme Scale Customer Scenarios and Use Cases
WSI35 - WebSphere Extreme Scale Customer Scenarios and Use CasesWSI35 - WebSphere Extreme Scale Customer Scenarios and Use Cases
WSI35 - WebSphere Extreme Scale Customer Scenarios and Use Cases
 
WSI33 - Advanced Performance Tactics for IBM WebSphere Application Server
WSI33 - Advanced Performance Tactics for IBM WebSphere Application ServerWSI33 - Advanced Performance Tactics for IBM WebSphere Application Server
WSI33 - Advanced Performance Tactics for IBM WebSphere Application Server
 
WSI32 - IBM WebSphere Performance Fundamentals
WSI32 - IBM WebSphere Performance FundamentalsWSI32 - IBM WebSphere Performance Fundamentals
WSI32 - IBM WebSphere Performance Fundamentals
 
W23 - Advanced Performance Tactics for WebSphere Performance
W23 - Advanced Performance Tactics for WebSphere PerformanceW23 - Advanced Performance Tactics for WebSphere Performance
W23 - Advanced Performance Tactics for WebSphere Performance
 
2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...
2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...
2689 - Exploring IBM PureApplication System and IBM Workload Deployer Best Pr...
 
2596 - Integrating PureApplication System Into Your Network
2596 - Integrating PureApplication System Into Your Network2596 - Integrating PureApplication System Into Your Network
2596 - Integrating PureApplication System Into Your Network
 
1457 - Reviewing Experiences from the PureExperience Program
1457 - Reviewing Experiences from the PureExperience Program1457 - Reviewing Experiences from the PureExperience Program
1457 - Reviewing Experiences from the PureExperience Program
 
CSD-2881 - Achieving System Production Readiness for IBM PureApplication System
CSD-2881 - Achieving System Production Readiness for IBM PureApplication SystemCSD-2881 - Achieving System Production Readiness for IBM PureApplication System
CSD-2881 - Achieving System Production Readiness for IBM PureApplication System
 
ACU-1445 - Bringing workloads into production on PureApplication System
ACU-1445 - Bringing workloads into production on PureApplication SystemACU-1445 - Bringing workloads into production on PureApplication System
ACU-1445 - Bringing workloads into production on PureApplication System
 
CIT-2697 - Customer Success Stories with IBM PureApplication System
CIT-2697 - Customer Success Stories with IBM PureApplication SystemCIT-2697 - Customer Success Stories with IBM PureApplication System
CIT-2697 - Customer Success Stories with IBM PureApplication System
 
CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...
CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...
CIN-2650 - Cloud adoption! Enforcer to transform your organization around peo...
 
IC6284A - The Art of Choosing the Best Cloud Solution
IC6284A - The Art of Choosing the Best Cloud SolutionIC6284A - The Art of Choosing the Best Cloud Solution
IC6284A - The Art of Choosing the Best Cloud Solution
 
C219 - Docker and PureApplication Patterns: Better Together
C219 - Docker and PureApplication Patterns: Better TogetherC219 - Docker and PureApplication Patterns: Better Together
C219 - Docker and PureApplication Patterns: Better Together
 
PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...
PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...
PAD-3126 - Evolving the DevOps Organization around IBM PureApplication System...
 
C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...
C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...
C418 - Build, Deploy and Manage Your First Open Pattern with PureApplication ...
 
7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local System
7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local System7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local System
7450A - CRONOS helping ENGIE adopting Private Cloud with Bluemix Local System
 
IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...
IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...
IBM Cloud University 2017 session BLUE010 - How Dutch Tax Built Their Core Bu...
 
IBM Think 2019 session 2116 - Best practices for operating and managing a pro...
IBM Think 2019 session 2116 - Best practices for operating and managing a pro...IBM Think 2019 session 2116 - Best practices for operating and managing a pro...
IBM Think 2019 session 2116 - Best practices for operating and managing a pro...
 

Recently uploaded

08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking MenDelhi Call girls
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxKatpro Technologies
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machinePadma Pradeep
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptxLBM Solutions
 
Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Paola De la Torre
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationRidwan Fadjar
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticscarlostorres15106
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationRadu Cotescu
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdfhans926745
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)Gabriella Davis
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
 
Slack Application Development 101 Slides
Slack Application Development 101 SlidesSlack Application Development 101 Slides
Slack Application Development 101 Slidespraypatel2
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsMemoori
 
Benefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksBenefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksSoftradix Technologies
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountPuma Security, LLC
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesSinan KOZAK
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024Rafal Los
 

Recently uploaded (20)

08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machine
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
Key Features Of Token Development (1).pptx
Key  Features Of Token  Development (1).pptxKey  Features Of Token  Development (1).pptx
Key Features Of Token Development (1).pptx
 
Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101Salesforce Community Group Quito, Salesforce 101
Salesforce Community Group Quito, Salesforce 101
 
My Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 PresentationMy Hashitalk Indonesia April 2024 Presentation
My Hashitalk Indonesia April 2024 Presentation
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmaticsKotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
Kotlin Multiplatform & Compose Multiplatform - Starter kit for pragmatics
 
Scaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organizationScaling API-first – The story of a global engineering organization
Scaling API-first – The story of a global engineering organization
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
Slack Application Development 101 Slides
Slack Application Development 101 SlidesSlack Application Development 101 Slides
Slack Application Development 101 Slides
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial Buildings
 
Benefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other FrameworksBenefits Of Flutter Compared To Other Frameworks
Benefits Of Flutter Compared To Other Frameworks
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path Mount
 
Unblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen FramesUnblocking The Main Thread Solving ANRs and Frozen Frames
Unblocking The Main Thread Solving ANRs and Frozen Frames
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 

W22 - WebSphere Performance for Multicore and Virtualised Platforms

  • 1. © 2010 IBM Corporation Conference materials may not be reproduced in whole or in part without the prior written permission of IBM. WebSphere Performance for MulticoreWebSphere Performance for Multicore and Virtualised Platformsand Virtualised Platforms Session Number: W22Session Number: W22 Hendrik van RunHendrik van Run –– hvanrun@uk.ibm.comhvanrun@uk.ibm.com
  • 2. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 2 Agenda • Multicore processor technology Why do we need/have multicore processors? Examples of Multicore processor architectures How to get value out of Multicore platforms? • Known Issues and Best Practices for Multicore Administration Programming • Virtualisation in the Enterprise Why Virtualisation? Different forms of Virtualisation Where to start? • Known Issues and Best Practices for Virtualisation Performance Overhead Resource Over-Commitment WebSphere Performance Best Practices Other Best Practices
  • 3. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 3 Why do we need/have multicore processors? Moore’s law • Moore's law describes a long- term trend in the history of computing hardware The number of transistors that can be placed (inexpensively) on an integrated circuit has doubled approximately every two years The trend has continued for more than half a century This is not expected to stop until 2015 or later • For almost two decades, single thread performance has doubled every 2 years in a very predictable fashion
  • 4. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 4 Why do we need/have multicore processors? CPU clockspeed has hit a wall • Performance has been propelled forward by continuing advancements in Clock speed Instruction Level Parallelism (ILP) L1/L2/L3 cache size • Heat dissipation increases linearly with processor frequency This is the main reason that clock speeds cannot easily exceed 3-5 GHz • Using mutiple cores on the same integrated circuit continues to allow for increased performance
  • 5. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 5 Why do we need/have multicore processors? The multicore future: System on a chip dreams • Massively complex and interesting problems await chip designers • How else can we use all these extra transistors? More processors on a chip (Multi-core) More threads on a core (SMT) Special-purpose processors (eg. crypto, compression) Integrate system components on the chip (eg. network, I/O) • Multicore is only the first, most notable impact of the physical phenomena of ever shrinking die sizes • The real story is that the measure of success for future processors is changing It’s no longer about being the Gigahertz champion or having the most clever microarchitecture It’s about who can use their transistor budget most effectively to speed up real workloads • One size will not fit all Greater focus is required on current and emerging workload characteristics Both general purpose and special purpose processing will play a role Processor and system effectiveness will be measured more aggressively against multiple criteria (performance, power, price, reliability) Greater levels of integration in hardware will drive new requirements into software
  • 6. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 6 Why do we need/have multicore processors? What are the benefits of multicore platforms? • Consolidation Most admins are looking at the sheer number of cores in a single box now as a means to consolidate older larger systems Allows for collapsing down of multi-tiered architectures • Power Savings Power costs in a system come from more than just the processor. Ability to consolidate down 4 machines into 1 will save you power. Eliminating machines from your infrastructure provides consolidation of the sheer amount of moving parts in your infrastructure saving you money. • Performance Moving from old single core hardware to new multicore hardware can result in significant performance improvements to legacy applications • All of the above effectively reduce cost
  • 7. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 7 Why do we need/have multicore processors? Why should I understand multicore platforms? • Typical hardware replacements cycles mean boxes with these processors are coming into your IT shop soon if not already there • These systems present unique challenges for Operating System, WebSphere, and Application Deployment outside of the past 10 years of knowledge • Hardware specific features and characteristics can cause major performance swings in applications • Requires developers to understand how to write code optimized for deployments on this type of hardware or/and Administrators that know how to work around multicore code problems. • Tighter coupling of operations and development is required • Through understanding of 3rd party libraries • Possible reworking of existing applications required to make them run optimally on micropartitioned or over committed hardware • Intelligent and insightful planning will ease your pain
  • 8. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 8 Why do we need/have multicore processors? Examples of Multicore processor architectures Intel Nehalem • 1.86 – 3.6 GHz • 2 – 8 cores per chip • Up to 2 threads per core (SMT) IBM Power7 • 3.0 – 4.25 GHz • 4 – 8 cores per chip • Up to 4 threads per core (SMT) IBM Power6 • 3.5 – 5.0 GHz • 2 cores per chip • Up to 2 threads per core (SMT) AMD Phenom II • 2.5 – 3.5 GHz • 2 – 6 cores per chip • Up to 2 threads per core (SMT) Sun Niagara 2 • 1.2 – 1.6 GHz • 4 – 8 cores per chip • Up to 8 threads per core (SMT)
  • 9. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 9 How to get value out of Multicore platforms? Virtualisation • Large number of cores can be a challenge How can you make sure that you’re using them all? • Virtualisation enables you to effectively run more applications on the same machine Different types of virtualisation exist We will discuss those in more detail towards the end of this presentation • There is no single “best” virtualisation technology out there Depends very much on your business and application needs • Start with requirements Do you want movable OS images? Do you need OS level partitioning? Do you need Dynamic resource allocation? Do you want to simply pay the lowest cost?
  • 10. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 10 How to get value out of Multicore platforms? Understand scaling of WebSphere Application Server • The evolution of multicore hardware places unique demands on application server performance A single instance must perform well at a low number of cores, but also scale to take advantage of many cores Multiple instances must be able to run concurrently providing “vertical scaling” on a single server • WAS V7 Multicore Performance WAS V7 achieves close to 1000 requests/second on a single CPU core running the DayTrader benchmark in this configuration WAS is able to take advantage of all 12 cores showing significant performance increases as more cores are enabled The ability to achieve high scalability depends on application architecture • WAS V7 Multicore Scalability The 2nd graph depicts WAS V7 scalability relative to a linear scale. With this benchmark, WAS V7 scales well to 8 cores. Scalability between 8 and 12 cores is beginning to flatten, but significant improvement is still being measured at 12 cores • WAS V7 Vertical Scaling WAS V7 ND provides the foundation and features for developing a vertically clustered application (or applications) on a single platform Vertical scaling can take advantage of many CPU cores and dramatically improve single server scalability
  • 11. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 11 Known Issues and Best Practices for Multicore
  • 12. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 12 Known Issues and Best Practices for Multicore Introduction • Multicore platforms require specific knowledge Administrators Developers Architects • Single thread application performance is typically lower Most Multicore processors run at a relative low clockspeed This fundamental limitation is central to what we will discuss here
  • 13. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 13 Known Issues and Best Practices for Multicore Administration • Multicore processor designs focus on building more capacity No longer an increase in clockspeed with every new processor • Clock speeds are up against the boundary of what’s possible Effects anything that is single threaded: Application Startup Time (i.e. 60 s instead of 15 s) WSAdmin and scripting runtime startup Batch compilation of JSPs Client Response Times (i.e. 1.5 s instead of 500 ms) • Hence you can no longer expect performance to double every 2 years Incremental improvements in processor design will be made optimizing pipelines and instruction cycles Unlikely to see similar increase from the past in single thread performance
  • 14. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 14 Known Issues and Best Practices for Multicore Administration • Memory capacity growth is slower than CPU capacity growth Most of the new multicore systems are limited in memory Forced to run 8-16 hardware threads or more against 8GB-16GB of space Be very careful not to overcommit your memory resources! 5% CPU utilisation and 100% memory consumption is common • Very advanced and sensitive BIOS and other hardware settings required to get the most performance SMT/CMT/HT effect application performance Intel/AMD “Turbo Mode” has drastic effects on different applications
  • 15. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 15 Known Issues and Best Practices for Multicore Administration • Consolidated environments can lead to issues with High Availability Less machines means a single outage results in more applications effected Need to plan to ensure data replication occurs off a single machine • Legacy software can suffer massively Massive concurrent load on code not designed or tested through enough to handle it Clock speed drop effects CPU intensive code paths • Hardware while becoming much more powerful is becoming more specialized
  • 16. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 16 Known Issues and Best Practices for Multicore Administration • WAS administrative best practices With previously discussed memory and CPU speed concerns Ensure your heaps and thread pools are correctly sized With WSAdmin scripts limit the initialization of the framework Ensure performance acceptance testing is completed on identical deployment topology Business level application planning should take into consideration hardware variations Work with developers to institute codepath focused profiling and improvement in applications • Garbage Collection Overhead Due to the increase in processing speed and capability on a single system the amount of garbage created in a single JVM increases and you must tune appropriately
  • 17. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 17 Known Issues and Best Practices for Multicore Programming • Programming for a multicore world is not THAT much different than writing good performing code today. Developers just need to do a little more due diligence before committing code Develop with the deployment hardware in the back of your head. You cannot treat the hardware as an unknown • Maximize single thread (request) performance Minimizes response time for end user requests Code for a request that can execute as efficiently as possible on a single thread leads to better performance and in most cases better scalability as concurrent requests go up Developers have to be more aware of what their code is doing as it flows through 3rd party libraries What to do: Focus on requests traversal from the socket down to the backend system Review your code to remove unnecessary caching or CPU intensive method calls
  • 18. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 18 Known Issues and Best Practices for Multicore Programming • Utilize Effective Memory Management The JVM for all of its advancements still has its performance dictated to it by garbage collection Minimize Object creation by avoiding the overly object oriented scenario as well as minimizing reflection. Keep it primitive where ever possible. Audit your code for calls to “new” but avoid pooling except for excessively complex object trees. If exploiting internal pooling or caching ensure it happens as high up the code stack as possible. Keep object structures as “thin” as possible. • Exploit Advanced Synchronization To be able to scale up effectively with the move to multicore hardware every developer needs to be well aware of explicit and implicit synchronization Define early on what data need to be shared across hardware threads and what doesn’t and use appropriate data structures. Ensure synchronization is at the lowest possible level. Decompose locking schemes. For shared data structures utilize java.util.conncurrent or advanced libraries Expose outwardly via java docs the level of synchronization to API level calls to your component for internal consumption.
  • 19. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 19 Known Issues and Best Practices for Multicore Programming • Exploit Effective Use of Debugging, Trace, and Logging Code Minimizes runtime codepath length and GC overhead The JIT is excellent at optimizing code out but it is still prudent to create clean code it is also critical to create performant trace code What to do: Create trace code for your method entry and exits but ensure they are properly gated Review your trace code to make sure it is putting out information to help with debugging but nothing more to keep overhead as low as possible • Remove unused or redundant code Reduces startup time by cutting the amount of class size being loaded from disk Redundant code and unused code contributes to class file size growth resulting in classload times increases What to do: Audit your code to ensure that all the method calls in your classes are actually accessed from other code, if not remove them. Ensure that in your methods all code is reachable by running one of the various Eclipse plugins that detects dead and unreachable code.
  • 20. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 20 Virtualisation in the Enterprise
  • 21. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 21 Why Virtualisation? • Most datacenters are heavily underutilized with systems running between 5-10% CPU on average This is not intentional but typically due to the unfounded need for all applications to run on discrete machines OR applications cyclical life cycles All modern hardware can easily run at above 50% CPU without losing performance • Not running your servers at maximum capacity is costing you dollars in terms of software licenses, power consumption, floor space, etc • There is no single solitary way to virtualize your environment. You need a strategy based off of your applications and business needs
  • 22. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 22 Why Virtualisation? • Consolidation Most admins are looking at the sheer number of cores in a single box now as a means to consolidate older larger systems Allows for collapsing down of multi-tiered architectures • Power Savings Power costs in a system come from more than just the processor. Ability to consolidate down 4 machines into 1 will save you power. Eliminating machines from your infrastructure provides consolidation of the sheer amount of moving parts in your infrastructure saving you money. • Performance Moving from old single core hardware to new multicore hardware can result in significant performance improvements to legacy applications
  • 23. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 23 Why Virtualisation? • Typical hardware replacements cycles mean boxes with these processors are coming into your IT shop soon if not already there so take advantage of them!! • Virtualization can drastically cut down on the energy and effort to get new WAS environments stood up for testing and production • In the world of every tighter budgets and constrained resources the ability to get the most out of your hardware makes a difference on the bottom line.
  • 24. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 24 Different forms of Virtualisation Hardware Virtualization • Advantages – Pure electronic isolation – Hardware based resource sharing – Higher Performance • Disadvantages – Management and maintenance of large topology of systems Software Virtualization OS Virtualization Power7 Hardware • Advantages – Hardware agnostic moveable OS images – OS based isolation – Software based resource sharing – Good Performance • Disadvantages – I/O system performance – System Tooling not accurate • Advantages – Single OS installation – OS isolation/security – Simplistic box based administration – Near on metal performance • Disadvantages – System level only management – Dependent on base LPAR/Hardware for availability
  • 25. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 25 Different forms of Virtualisation Multiple WebSphere Instances Single WAS multiple applications WAS App App App App App App App App • Advantages – Single OS installation – Process based security – Fast install, setup and boot – Best performing setup – Least memory consumption • Disadvantages – Process level security only • Advantages – Single OS installation – Simplistic Administration – Good performance • Disadvantages – No application isolation – Rouge Applications – Minimal availability
  • 26. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 26 Where to start with Virtualisation? • Start small and work big towards a grander plan • By starting small and slowly rolling applications from “bare metal” into your virtualized infrastructure you can work through issues and adjust future plans • By having a grand vision on where the datacenter in your business is heading its easier to sell the long term value and show short term progress in real numbers • Begin by working with back office, old non critical, new applications, or development • New applications No known baseline to be compared to so no one can say “the old stuff was faster” Can be build and tested from the beginning with virtualization in mind. Thus all performance testing, etc can be done on like systems. • Back office / Old non critical applications Nice starting point as huge amount of information on access patterns and other factors affecting performance exist. Minimal risk of impacting large user sets • Gain knowledge and roll that knowledge into your long term plan
  • 27. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 27 Known Issues and Best Practices for Virtualisation
  • 28. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 28 Known Issues and Best Practices for Virtualisation Virtualisation Performance Overhead • Performance overhead differs by platform and virtualization technology • Completely virtualized environments typically have higher performance penalties Network and disk IO latencies significantly higher in these environments • Para-virtualization technologies provide an optimized path to the underlying hardware for these high latency paths VMware Tools provide this capability in VMware environments Optimizations for virtual Ethernet driver, memory sharing and time sync Enabling VMI kernel support on VMware/Linux-based guests reduces virtualization overhead further Check VMware documentation for supported kernels • Newer hardware technologies provide better virtualization support and performance For example: Intel VT and AMD-V
  • 29. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 29 Known Issues and Best Practices for Virtualisation Virtualisation Performance Overhead – VMware ESX 3.5 • Example overhead measurements based on DayTrader application Minimal disk I/O characteristics Increased I/O activity will increase performance delta • VMI kernel reduces overhead further (~ 5% overhead)
  • 30. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 30 Known Issues and Best Practices for Virtualisation Virtualisation Performance Overhead – PowerVM LPARs • Traditional “Thick” LPARs provide hard isolation and dedicated resources “Thin” LPARs use the AIX Network Installation Manager (NIM) to share a single remote OS boot image • Minimal disk I/O characteristics are behind the small performance delta
  • 31. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 31 Known Issues and Best Practices for Virtualisation Virtualisation Performance Overhead – PowerVM WPARs • Workload Partitions (WPARs) provide a virtualized operating environment activated within the global OS • Virtualization overhead minimal • Multi-instance scalability impacted by normal process scheduling hurdles RSET can be used to achieve optimal throughput for multiple instances in both LPAR and WPAR environments
  • 32. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 32 Known Issues and Best Practices for Virtualisation Virtualisation Performance Overhead – Different Levels of Isolation
  • 33. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 33 Known Issues and Best Practices for Virtualisation Resource Over-Commitment • Memory Never over commit physical memory in the system if you expect predictable performance When pushing the boundaries of physical RAM in a system smaller JVM heaps allow for more WAS instances on a single hardware machine • CPU CPU over commit results in good performance up to the point of 100% saturation on the hardware At which point, response time begin to suffer Limit Memory Over commit Lower CPU Utilization Over commit
  • 34. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 34 Known Issues and Best Practices for Virtualisation WebSphere Performance Best Practices • Thread pool settings Must be sized not to the number of virtual CPUs exposed but to the number of physical CPUs on the system • Heap Sizes Set heap sizes to ensure that guest remain in physical memory for optimal performance. Right size your heap to ensure you are not wasting memory • High Availability You must implement a intelligent placement mechanism to ensure fault tolerance for your applications
  • 35. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 35 Known Issues and Best Practices for Virtualisation WebSphere Performance Best Practices • Deployment Scripting Script your administrative tasks to ensure that installation and configuration of multiple nodes is repeatable Allows you to bundle and distribute images quickly and have them up and running in a “certified” manner • Minimizing IO Minimize the application server IO to ensure maximum performance Cut logging to a bare minimum Bare minimum writing to disk should be done unless the disk system is optimized Ensure “chattiness” of your application is as minimal as possible
  • 36. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 36 Known Issues and Best Practices for Virtualisation Other Best Practices • Resource utilization monitoring Virtualization often skews the CPU, IO and other resources metrics reported by the Guest OS Need to rely on the Hypervisor for resource utilization statics Example: AIX micro-partitioning • Clock sync / drift Problem noticed in VMware / WebSphere HA environments Installation of VMware Tools and use of VMI kernel address this issue • VM migration tools (ie. VMotion) Avoid these technologies in your environment
  • 37. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 37 Known Issues and Best Practices for Virtualisation Other Best Practices • WebSphere tuning can be directly linked to resources allocated to Guest VM • Reduce disk space and memory requirements for Guests by disabling or removing un-used services and applications • Leverage VM cloning and templating technologies to provide a uniform operating environment For example: “Linked-clone Templates” in VMware Scripts for changing WAS hostnames http://www.ibm.com/developerworks/websphere/techjournal/0905_webcon/0 905_webcon.html • Use SANs or other fast disk technologies for IO intensive workloads • Can I trust Java profiling tools in virtualized environments? Successfully used Java Health Center CPU profiling to diagnose performance issue
  • 38. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 38 Deploy Applications in Virtualised Environments WebSphere CloudBurst 2.0 Appliance • Ability to rapidly deploy IBM middleware and customer applications into virtual and cloud environments leveraging best practice deployment patterns • Support for multiple hypervisor platforms, such as VMware ESX, PowerVM, and z/VM 38 • Availability of the following product images WebSphere Application Server Hypervisor Edition V6.1 and V7 WebSphere Application Server Hypervisor Edition V7 now available on Red Hat Enterprise Linux! WebSphere Application Server Hypervisor Edition - Intelligent Management Pack Available for WebSphere Application Server Hypervisor Edition V7 WebSphere Process Server Hypervisor Edition V6.2 and V7
  • 39. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 39 Summary • Multicore processor technology and Virtualisation Technologies are here today And this trend will continue… • WebSphere performs well on those platforms But make sure to address known issues and apply best practices
  • 40. WebSphere Technical Conference and Portal Excellence Conference © 2010 IBM Corporation 40 QuestionsQuestions • Please remember to complete your session evaluation form Please complete your session evaluation, session number is W22