Technical detail

The IBM® WebSphere® DataPower Low Latency Appliance XM70 extends your ESB’s capability by supporting the following functionality:

• Front-side RMM/RUM receiver
  - Multicast and Unicast
  * Reliable Multicast Messaging (RMM)
  * Reliable Unicast Messaging (RUM)
  * Reliable UDP (RMM with ACK)
  - Peer-to-peer support
  - One-to-many / many-to-many support
  - Filters, message selectors, turboflow support
  * Both topic/queue and message-based filtering
  - “fastpath” – no multistep
  - Integration with multistep processing
  * Content-based routing
  * AAA (Authentication, Authorization and Audit)
  * Any-to-any transformation
  * Protocol bridging

• Backside RMM/RUM transmitter

• Quality of Service
  - Total message ordering for both input and output
  - Congestion management
  - Horizontal and vertical scalability
  * Load distribution
  * Best practices configuration

• Reliable Multicast using UDP
  - Highly scalable, topic-oriented communication
  - Supports PGM specification.
  - NAK reliability
  - Sliding history window
  - Congestion management

• Reliable Unicast using UDP
  - Efficient point-to-point, topic-oriented communication.
  - Higher reliability similar to that of TCP
  - ACK reliability
  - Periodic heartbeat ensures liveness

• Reliable Unicast Messaging
  - High throughput, topic-oriented communication using TCP
  - Simpler than TCP streaming primitives
  - Periodic heartbeat ensures liveness
  - Can be used in WAN or through firewall

• Protocol bridging
  - MQ LLM, MQ,TIBCO Rendezvous, TIBCO EMS, WebSphere JMS, HTTP(S)

• High-speed content-based routing
  - FIX over MQ LLM transport
  - XML over MQ LLM transport

• Any-to-any data transformation (there will be a latency penalty)

• Web Services Gateway

• High availability – fault tolerant configuration of 2 or more devices
  - Tiered topics and queues across multiple devices
  - Primary and secondary transmitters / stream failover

• Load distribution across devices for workload management
  - Options
  * Protocol level load distribution
  * Configuration and deployment-driven best practices
  - Needs to be combined with fault tolerant deployment and configuration as well

• Total message ordering across devices
  - Total order receiver
  - Transmitter synchronization

• Congestion management

• Message persistence for failover, replay and late join
  - In-memory persistence is core requirement
  - Need to look at disk persistence as well across device restarts