Informatica’s Ultra Messaging™ (UM) is a pioneering, high-speed middleware. Many contemporary middleware systems have a central routing and monitoring component known as a ‘daemon’
For performance reasons UM does not use a central daemon, thus removing a significant bottleneck and gaining substantial performance benefits. Yet, how can operations teams monitor this environment?
On-the-wire monitoring techniques can measure message latency, but cannot examine the internal state of the middleware. The Geneos UM monitoring interface, accesses the monitoring statistics available from UM-based application components, the UM Persistence (UMP) store and the UM Gateway to provide visibility of the underlying factors contributing to operational issues and message latency. This is particularly useful in predicting impending issues and assisting with the rapid diagnosis of current issues, significantly reducing the time to identify faults. Geneos combines the UM middleware metrics with other data metrics acquired across the trading infrastructure, for example round-trip message latency and interpretation of FIX message flows to provide an end-to-end view of order flow performance. Geneos can also go further and enable operations teams to set up customisable rules and alerts to correlate events, potentially avoiding critical trading disruption.
The Geneos UM monitoring interface provides managers with a total view of the technology supporting transaction flows, including all UM transport types and non-transport components such as the UM Store and UM Gateway. It delivers numerous benefits to your business and reduces the time for diagnostic analysis from hours to minutes. It helps you mitigate risk by accelerating intervention strategies and identifying unusual application behaviour fast. The interface helps you to improve performance by providing visibility of factors contributing to application latency not just an indication of the actual latency. It also collects historical UM for further analysis and to inform Geneos rule configuration.
Operations staff
Incorporates multiple UM domains into an overall Geneos-managed monitoring infrastructure
Comprehensive support for all UM statistics and components
Highlights key metrics across the infrastructure
Exposes problem components
Develops procurement and technology upgrade plans based on historical performance data
Development staff
Provides in-depth statistics on component operation and performance
Assists development by revealing consequences of design decisions
Shows efficiency of transport-session configurations
Provides tangible evidence of performance of pre-live testing
Supports transmission control protocol (TCP), reliable multicast, unicast, shared memory (IPC) and InfiniBand™-based communication
Enriches Ultra Messaging™-provided statistics with rates and ratios
Provides a central monitoring point across topic-resolution domains
Optimises middleware configuration
Early warnings of impending operational and performance issues
Supports Latency Busting Middleware (LBM) versions 4.1 onwards
The plug-in translates and enriches monitoring statistics generated by UM Store-based components into Geneos views. Ultra Messaging can be made to generate such statistics by configuration options only – no changes to applications are necessary. The plug-in also supports statistics from theUMP store process and the UM Gateway, which are ordinarily only available via a browser and must be updated manually.
A minimum of one UM plug-in is required to monitor each UM topic-resolution domain (two if resilience is required). Using this configuration, Geneos rules can be configured on each view alerting the operations team to failures and other operationally significant circumstances.
The UM interface provides a clear picture of every aspect of Ultra Messaging that contributes to correct operation and performance. At the same time, no changes are required to the underlying application (monitoring statistics generation is enabled and disabled using configuration parameters).
The UM highlights the numbers and rate of irretrievable message loss, negative acknowledgements (NAKs) generated and messages recovered. It also reports on the number of sources and topics being persisted on each store, and the number of topics and messages being passed between domains using the UM Gateway.
Geneos is able to capture and aggregate this data in the Geneos Gateway at a granular level, right down to microsecond time-stamping from the underlying server platforms. ITRS Geneos users can go even further by creating their own customised views of the data gathered and maintained by Geneos Gateways.
Context views: top-level statistics on each UM-based application.
Monitoring the success of the topic-resolution process.
Event queue views: detailed statistics on UM event queues.
Monitoring the delays introduced within application components.
Source and receiver views: monitoring the performance of each source and receiver in an application.
Identifying the busiest sources and receivers.
Identifying receivers causing performance bottlenecks.
Monitoring the efficiency of the transport configuration.
UM Store views: statistics on the operation of UM stores.
Reconciling topics and sources being persisted.
UM Gateway views: statistics on the operation of UM Gateway peer and endpoint portals.
Verifying correct gateway operation.
Diagnosing gateway operational and performance issues.
For more information on the Ultra Messaging plug-infor Geneos