Service message system for a switching architecture

Abstract

A service message system for a switching architecture including at least one Switch Fabric (10, 20) comprising a switch core (15, 25) located in a centralized building and a set of Switch Core Access Layer (SCAL) elements distributed in different physical areas for the attachment to the different Port adapters (30, 31). Each SCAL elements particularly includes a SCAL receive element (11-i) and a SCAL Xmit element (12-i) for the respective access to one input port and one output port via serial links. The service message is based on the use of a Cell qualifier field at the beginning of each cell, which comprises a first and a second field. The first field is the Filtering Control field which permits an easy decoding of a service message cell, when applicable. The second field is used for determining which particular type of service message is conveyed via the cell. Following the Cell qualifier is the Switch Routing Header (SRH) which permits the characterization of the destination of the cell and is used for controlling the routing process. Preferably, the service message is used in a fault tolerance configuration where two different Switch Fabrics act as a standby to each other and shares a part of the traffic. Each one is configured as a default routing path for some ports adapters and a backup path for the others. In that particular configuration, the service message system of the invention uses the first field of the Cell qualifier to transport a Direct filtering command causing the Switch fabric to route the cell when the SRH is representative of its default output port destination. Conversely, the first field may transport a Reverse filtering command in the first field that causes the Switch fabric to reverse the default routing process. The first field is also used for characterizing a service message cell which the second field indicates the accurate type. Particularly, two types are used for the production of the filling cells when no data cell is to be transmitted at a particular location of the switching architecture.

Dependencies are inferred from the claim text.