UME Key Concepts

A deployed UME system comprises multiple instances of sources, receivers and persistent stores under the control of one or more applications. UME's reliance on LBM's innovative messaging model and its own Parallel Persistence® sends high message throughput between all components with minimal latency. With the highly configurable Persistent Store, a UME system can also easily recover from the failure of any number of components.

Persistence Designed for Speed and Effortless Recovery

  • Stores and Receivers. UME stores do not sit in the critical message path to receivers. The UME design runs stores and receivers in parallel, avoiding the bottlenecks and latencies of legacy store-and-forward designs.
  • With the UME Persistent Store message streams live on even after the source sending them has gone away. Further, a source need not wait until a receiver is ready before sending messages. In short, sources and receivers need not be running at the same time in order to pass messages.
  • Durable Subscriptions. A receiver can stop and restart with no loss of messages, even if new messages were sent while the receiver was stopped. The durable source is also beneficial in that a source can stop and restart without receivers ever knowing.
  • UME implementations feature no single point of failure. Should the primary store become inactive, a secondary store assumes the role of the primary store.

Sources and Stores: A Partnership for Persistence

  • Stability notifications. Sources retain all messages until notified by the active store(s) that they are stable. If a store fails, the source can use retained messages to bring the store up to date upon restart. Stability notices flow in parallel to the message stream as well, adding to UME's performance advantage over legacy persistent messaging systems.
  • Since memory and disk storage are finite, UME allows administrators to configure source retention policies. For any source, the administrator can configure a size limit for total retained messages and a threshold below which no retained messages are released.
  • With LBM's straightforward configuration options, sources can be configured to use redundant stores in either a round robin or quorum/consensus manner. In a round robin configuration, a source retains messages until they are considered stable at the active store. In the case of quorum/consensus configuration, a message is considered stable after it has been successfully stored within a group of stores or among groups of stores.
  • Quorum/consensus store usage also allows groups of stores to be configured across multiple locations to ensure continued operation in the event of a site-wide failure at any location.
  • A UME source can also be configured to release a retained message when it receives a delivery confirmation event from the receiver. UME receivers can decide to generate the confirmation event when the message is passed to the receiving application, or when the receiving application releases it.
  • Late Join. UME provides a late join feature that allows a new receiver to join a group of receivers already listening to a source and receive messages that were sent before it joined the group. This feature, also available in LBM, permits the source to limit the range of messages available to the late-joining receiver via its message retention policy.

Configure the UME Persistent Store for Your Critical Path

The UME Persistent Store presents multiple configuration options that adapt the store(s) to your messaging applications and hardware capabilities.

  • You can configure multiple UME store processes with one or more stores in each. Each store can have its own attributes and one or more topics. For the least impact after a store process failure, configure one store process for each topic.
  • To distribute the topic processing load, you could use one store for each topic, all running under one store process. For capable hardware where processing doesn't need to be distributed, you can configure all topics in one store running in one store process.
  • A store can persist one topic, more than one topic or persist all topics that match a specified pattern.
  • Persisted topics or topic patterns can be configured with separate options such as using a memory store or disk store. Short lived topics needing greater speed can be persisted on a memory store. Topics higher on the reliability scale can be persisted on a disk store.
  • Other topic options include repository size limits and message ages to help manage the volume of persisted messages.
  • You can run your UME Persistent Store(s) with a Web Monitor that provides a rich source of information about the health and activities of your store(s). The UME Web Monitor provides separate pages with statistics on the store, sources publishing on the topics being persisted at the store and the receivers listening on those topics.