Federation is one of the key elements of sense and ensures the overall system automatically establishes real-time relationships between all the discoverable sense nodes in a network, irrespective of the physical location or runtime infrastructure of the nodes themselves.

A sense node is a computational instance that is able to provide and manage all sense services, and to make them join virtually with others of the same service family installed on other parts of the network. This means that all services from the same family are visible to each other no matter where they are installed. Belonging to the same service family means being able to provide the same set of services to a requester, but perhaps using a different implementation or indeed perhaps a completely different service provider.

When joining a federation sense will monitor the status of the federation and the status of all the services within the federation by:

• monitoring each node’s perspective from all other nodes – which means no centralized controller or single-point-of-failure
• acquiring information regarding all the available protocols that the services on the nodes have made available – allowing sense to instruct intra-sense communication multiplexing on the most efficient protocol available at the time the communication is needed, again increasing system resilience

A federation snapshot is available through the sense web console and shows which nodes and services are running in the network and how they are performing. The federation is the unique place where information about services, feelings, business flows and nodes is stored together with their individual and collective ‘health’ status.

The sense federation enables services to compute on the same virtual space of computing resources and to be extended from one space to another.

sense federation

sense is able to define the federation within in a local network of a server farm(s) and to establish federation outside on other networks (such as clouds) or external  or virtual server farms. Federation through internet is done by configuring sense‘s nodes in a secure way, sense then starts managing these nodes availability and measuring network latency to let all the other services within the federation know which will be the most performant service instance (of the family) that can be called to accomplish work.

The federation is not continuously monitored by a federation daemon, which would create a lot overhead, but is only refreshed at each ‘life sign’ of the system when something is changing, either getting better or worse i.e.  event based refresh on any status change. _sense’s_ federation cache is held in memory for highest performance.  It is also possible to register the federation cache to an XML or relational database.

A cache for the cloud

sense federation establishes a distributed caching mechanism for the cloud to enable sharing of information and business object around the cloud while making information available near the computing platforms themselves. This removes the idea of having information concentrated in a controller server while sparsely distributed on the cloud. Each sense instance shares the cache and can implement logic for managing information in the cache; an example could be a feeling that caches the last 100 different users’ call to the cloud and the mock strategy that resurrects this cache if something goes drastically wrong.

By federating information in the cloud in this way, each sense node is transparently aware of all other sense nodes and services via a local cache, for performance, yet is current as of the last interesting event change.

go to faq