EP0110569A2
Expired / Lapsed
EP
⏱️ 0.0 years remaining
Multiprocessor multisystem communications network
Assignee: Tandem Computers Inc
•
Filing: 1983-10-28
•
Granted: —
•
Expires: 2003-10-28
Abstract
Abstract A plurality of multiprocessor systems (a, b ...n) is arranged in a high speed network to allow any processor (10) in one system to communicate with any processor (10) in another system. The network may be configured as a multi- node dual bidirectional ring having a multiprocessor system at each node (18, 20). Packets of information can be passed around the ring in either of two directions and are temporarily stored in buffer memory locations dedicated to a selected destination processor (10) in a selected direction between each successive transfer between neighbouring nodes (18, 20). The buffer locations are managed so that a node (18, 20) can request an adjacent node (18, 20) to stop transmitting packets if the buffer is becoming full from that direction and request resumption of transmission of packets as the buffer empties.
Inventors
- 👤 James Conard Allen
- 👤 Wendy Bohle Bartlett
- 👤 Hoke Stephens Johnson
- 👤 Steven Derek Fisher
- 👤 Richard Oliver Larson
- 👤 John Charles Peck
Bibliographic Details
Application Number
—
Jurisdiction
EP
Data Source
google_patents
Last Synchronized
2026-11-04 04:14
CPC Classifications
G; G06; G06F; G06F13/00; G06F15/00; G06F15/16; G06F15/163; G06F15/173; G06F15/17337
Claims Analysis
Total Claims: 4
Claim 1
Independent Claim
170 words
A network of multiprocessor systems characterised by: a plurality of node means (18,20) connected to one another via data lines (22,24,26,28), each of said node means (18,20) comprising an interface means connected to a multiprocessor system (a,b...n), each of said multiprocessor systems comprising a plurality of processors (10) communicating with an interprocessor bus (12,14); buffer memory means in each of said interface means for storing packets of information transferred to said interface means from the one of said multiprocessor systems to which it is connected and from adjacent node means in the network, each of said packets of information indicating that one of said processors in said network to which it is to be sent; and control means in said interface means for transferring said packets'of information between adjacent ones of said nodes (18,20) and to the one of said processors (10) for which it is destined from the one of said nodes connected to the one of said multiprocessor systems containing the processor for which said packet is destined.
Claim 2
Dependent Claim
170 words
A network of multiprocessor systems characterised by: a plurality of node means (18,20) connected to one another via data lines (22,24,26,28), each of said node means (18,20) comprising an interface means connected to a multiprocessor system (a,b...n), each of said multiprocessor systems comprising a plurality of processors (10) communicating with an interprocessor bus (12,14); buffer memory means in each of said interface means for storing packets of information transferred to said interface means from the one of said multiprocessor systems to which it is connected and from adjacent node means in the network, each of said packets of information indicating that one of said processors in said network to which it is to be sent; and control means in said interface means for transferring said packets'of information between adjacent ones of said nodes (18,20) and to the one of said processors (10) for which it is destined from the one of said nodes connected to the one of said multiprocessor systems containing the processor for which said packet is destined.
Claim 3
Dependent Claim
310 words
A network of multiprocessor systems (a,b...n), each including a plurality of processors (10) sharing an interprocessor bus (12,14), said network being characterised by: a plurality of nodes (18,20) connected to one another by data links (22,24,26,28), each of said multiprocessor systems directly communicating with at least one of said nodes; buffer memory means in each of said nodes for storing information to be transferred to individual ones of said processors in all multiprocessor systems in said network, said buffer memory means having a plurality of storage locations and being configured such that certain ones of said storage locations are associated with each individual processor (10) in said network; first interface means in each of said nodes connected to the interprocessor bus of one of said multiprocessor systems and communicating with said buffer memory means therein for transferring information between individual processors in that one of said multiprocessor systems and one of said storage locations in said buffer memory means associated with said individual processor (10); at least one second interface means in each of said nodes communicating with said buffer memory means therein and connected to one of said data links for transferring information into and out of said storage locations in said buffer memory means associated with the same one of said individual processors in adjacent ones of said nodes; control means in each of said nodes for accomplishing the transfer of information between one of said storage locations in buffer memory means in that node and one of said data links connecting that node to adjacent ones of said nodes via one of said first interface means; and control means in each of said nodes (18,20) for accomplishing the transfer of information between ones of said storage locations in buffer memory means in that node and individual ones of said processors associated therewith via said second interface means.
Claim 4
Dependent Claim
310 words
A network of multiprocessor systems (a,b...n), each including a plurality of processors (10) sharing an interprocessor bus (12,14), said network being characterised by: a plurality of nodes (18,20) connected to one another by data links (22,24,26,28), each of said multiprocessor systems directly communicating with at least one of said nodes; buffer memory means in each of said nodes for storing information to be transferred to individual ones of said processors in all multiprocessor systems in said network, said buffer memory means having a plurality of storage locations and being configured such that certain ones of said storage locations are associated with each individual processor (10) in said network; first interface means in each of said nodes connected to the interprocessor bus of one of said multiprocessor systems and communicating with said buffer memory means therein for transferring information between individual processors in that one of said multiprocessor systems and one of said storage locations in said buffer memory means associated with said individual processor (10); at least one second interface means in each of said nodes communicating with said buffer memory means therein and connected to one of said data links for transferring information into and out of said storage locations in said buffer memory means associated with the same one of said individual processors in adjacent ones of said nodes; control means in each of said nodes for accomplishing the transfer of information between one of said storage locations in buffer memory means in that node and one of said data links connecting that node to adjacent ones of said nodes via one of said first interface means; and control means in each of said nodes (18,20) for accomplishing the transfer of information between ones of said storage locations in buffer memory means in that node and individual ones of said processors associated therewith via said second interface means.
References Cited (Backward Citations)
No citation records indexed.
USPTO Annuity & Maintenance Windows
Maintenance schedule is calculated for granted US utility patents.