Memory-to-memory topology: Peer-to-peer function with a local replicator

The basic peer-to-peer (both client and server function, or both mode) topology is the default configuration. This configuration uses a local replicator (a replicator on the same server as the session manager) and no partitioning.

In this basic peer-to-peer topology, each server Java Virtual Machine (JVM):

This configuration represents the most consolidated topology, where the various system parts are collocated and requires the fewest server processes. When using this configuration, the most stable implementation is achieved when each node has equal capabilities (CPU, memory, and so on), and each handles the same amount of work.

This topology is the most redundant because everyone replicates to everyone. On any failure recovery scenario, the server routed to already has a copy of the session. But as you add servers, more overhead (both CPU and memory) is needed to deal with replication. Without Memory-to-memory session partitions, scaling breaks down the quickest with this configuration. When scaling breaks down because of the machine characteristics, the client hit rate, the network speed, and the size on average of the HTTP session. The remaining topologies: client/server with isolated replicators, client/server with remote replicators, and peer-to-peer with remote replicators help mitigate and manage the scaling characteristics of replication. Replication costs are also reduced through the use of partitioning and time based write (employed by default) in these topologies.

It is also important to note that when using the peer-to-peer topology, that one session replication backup must run at all times for invalidation to occur. For example, if you have a cluster of 2 application servers, server1 and server2, that are both configured in the peer-to-peer mode and server2 goes down. All of backup information for server1 is lost and those sessions are no longer invalidated properly.


Related concepts
Memory-to-memory replication
Memory-to-memory topology: Peer-to-peer function with remote/isolated replicators
Memory-to-memory topology: Client/server function with isolated replicators
Memory-to-memory topology: Client/server function with remote replicators
Related tasks
Configuring memory-to-memory replication for the peer-to-peer function with a local replicator (default memory-to-memory replication)
Configuring memory-to-memory replication for peer-to-peer functions with remote/isolated replicators
Configuring memory-to-memory replication for the client/server function using isolated replicators
Configuring memory-to-memory replication for the client/server function using remote replicators



Searchable topic ID:   cprsm2mp2pdefault
Last updated: Jun 21, 2007 9:56:50 PM CDT    WebSphere Application Server for z/OS, Version 5.0.2
http://publib.boulder.ibm.com/infocenter/wasinfo/index.jsp?topic=/com.ibm.websphere.zseries.doc/info/zseries/ae/cprs_m2m_p2p_default.html

Library | Support | Terms of Use | Feedback