IBM developerWorks: Wikis - WebSphere eXtreme Scale Wiki

If you log in with your developerWorks ID, you can leave comments and feedback for the development team.

Each week, Billy Newport discusses hot topics, answers questions and describes features of WebSphere eXtreme Scale. The episodes are available on iTunes and YouTube:

eXtreme Scale podcast episodes:

Episode 1 - 2/27/2009 - Preloading, duplicate data, dynamic map creation, and putAll().
Episode 2 - 3/5/2009 - L2 cache for JPA, write-behind and high availability.
Episode 3 - 3/13/2009 - JPALoaders, database synchronization, WAS integration, and JPA vs. ObjectGrid APIs.
Episode 4 - 3/20/2009 - Clearing a map, Spring integration, and zones.
Episode 5 - 3/27/2009 - Trial overview, heap based evcition, removeAll(), near cache and write-behind with sync replicas.
Episode 6 - 4/4/2009 - Embedded vs. external grid, 2PC, HTTP Session replication, and choosing partitions.
Episode 7 - 4/12/2009 - PureQuery loaders, stored procedure caching, stateful singletons and large query sets.
Episode 8 - 4/22/2009 - Eliminating parallel queries, performance testing, write-behind error handling, reverse indexes and Impact 2009 preview
Episode 9 - 5/2/2009 - IBM Impact 2009 session overview, classifying styles of XTP and hub and spoke replication
Episode 10 - 5/23/2009 - Impact wrap-up, routing, patterns and optimistic locking
Episode 11 - 6/27/2009 - Version 7 overview, memory density, ADO.NET REST Data Service

Episode 1 - 2/27/2009 - Preloading, duplicate data, dynamic map creation, and putAll().

Data preloading techniques; Loader plug-ins vs. Client loaders
Duplicating data over all partitions and denormalization
Dynamic map creation
ObjectMap.putAll behaviors

Episode 2 - 3/5/2009 - L2 cache for JPA, write-behind and high availability.

Using eXtreme Scale for a L2 cache for OpenJPA or Hibernate.
Write-behind database synchronization overview and use cases.
Achieving high availability with eXtreme Scale.

Episode 3 - 3/13/2009 - JPALoaders, database synchronization, WAS integration, and JPA vs. ObjectGrid APIs.

How to use JPALoaders for automatically integrating with relational databases.
How to deal with third party or legacy applications changing a database directly without updating the cache.
How WebSphere eXtreme Scale integrates with WebSphere Application Server.
Whether an application should use JPA APIs or WebSphere eXtreme Scale APIs to access its persistence data.

Episode 4 - 3/20/2009 - Clearing a map, Spring integration, and zones.

How to efficiently clear the contents of a map.
Overview of eXtreme Scale Spring integration features.
How to use eXtreme Scale zones to control primary and replica placement.

Episode 5 - 3/27/2009 - Trial overview, heap based evcition, removeAll(), near cache and write-behind with sync replicas.

An overview of the eXtreme Scale trial, including it's capabilities and limitations.
Using heap level based eviction with a traditional eviction strategy.
Using removeAll() as a high performance alternative to remove().
The near cache: What it is and when to use it.
Using write-behind with a synchronous replica: How does it work?

Episode 6 - 4/4/2009 - Embedded vs. external grid, 2PC, HTTP Session replication, and choosing partitions.

How to choose whether to embed the grid in application processes or run in separate processes.
Why eXtreme Scale doesn't provide 2-phase commit protocol (2PC).
Using eXtreme Scale for HTTP Session replication.
Best practice for choosing the number of partitions per process.

Episode 7 - 4/12/2009 - PureQuery loaders, stored procedure caching, stateful singletons and large query sets.

Using IBM Data Studio PureQuery to create a general purpose loader.
Using maps and loaders to cache stored procedure results.
Techniques for caching query results in blocks for pagination support.
How eXtreme Scale can manage singleton services with highly available state.

Episode 8 - 4/22/2009 - Eliminating parallel queries, performance testing, write-behind error handling, reverse indexes and Impact 2009 preview

Techniques for eliminating parallel, grid-wide queries.
Discussion on why warming-up a JVM is important when performance testing.
How to handle write-behind errors.
Using reverse indexes for fast lookups.
Brief preview of what the eXtreme Scale Team is presenting at Impact.

Episode 9 - 5/2/2009 - IBM Impact 2009 session overview, classifying styles of XTP and hub and spoke replication

Overview of eXtreme Transaction Processing (XTP).
How to choose between class 1 and 2 XTP systems.
Hub and spoke replicated transactional systems with eXtreme Scale.
Overview of eXtreme Scale 6.1.0.5 cumulative fix 2.
More details of the eXtreme Scale sessions at IBM Impact 2009.

Episode 10 - 5/23/2009 - Impact wrap-up, routing, patterns and optimistic locking

Recap of what happened at IBM Impact 2009 in regards to eXtreme Scale.
How eXtreme Scale routing works for clients.
How and why to use the numInitialContainers deployment policy attribute when starting large grids.
Overview of optimistic locking.
Patterns for integrating eXtreme Scale with applications.

Episode 11 - 6/27/2009 - Version 7 overview, memory density, ADO.NET REST Data Service

Overview of what's new in WebSphere eXtreme Scale V7.0
Memory density improvements in V7.0
How to lower back-end costs considerably using IBM WebSphere eXtreme Scale
Description of WebSphere eXtreme Scale chargeable units. What are they?
Overview of the upcoming ADO.NET REST Data Service.