[Proceedings] GLOBECOM '90: IEEE Global Telecommunications Conference and Exhibition
DOI: 10.1109/glocom.1990.116760
|View full text |Cite
|
Sign up to set email alerts
|

A multi-purpose memory switch LSI for ATM-based systems

Abstract: An TM switching system should be capable of handling multimedia traffic, which has different bit-rates and diversiied quality requirements. A self-routing switch is a key piece of equipment for attaining this goal. This paper proposes. a new switch construction and describes its LSI design and fabrication.The proposed switch, which is a memory switch with dynamic link speed control, can dynamically change link speeds according to traffic flow, thereby achieving high performance in any imbalanced traffic condit… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2

Citation Types

0
2
0

Publication Types

Select...
3
2

Relationship

0
5

Authors

Journals

citations
Cited by 10 publications
(2 citation statements)
references
References 2 publications
0
2
0
Order By: Relevance
“…The experimental ATM system developed by NTT [51] can be described as an L + M" + L single-route switching fabric: the L-switch at the outputs is not strictly necessary, from a functional point of view, but has been introduced to avoid the need of arbitrating the access of the various input interfaces to the pool of VCIs available at each output port, thus allowing the implementation of a fully distributed control architecture. The multiplex switch described in [51] can be considered at the top level as a 3-stage Clos arrangement, with module sizes 4 x 4, 64 x 64 and 4 X 4 respectively. The inner 64 x 64 modules are realized with parallel planes of banyan networks made of 4 x 4 elements: plane parallelism and an internal 4:l speed advantage make these modules nonblocking.…”
Section: Nttmentioning
confidence: 99%
See 1 more Smart Citation
“…The experimental ATM system developed by NTT [51] can be described as an L + M" + L single-route switching fabric: the L-switch at the outputs is not strictly necessary, from a functional point of view, but has been introduced to avoid the need of arbitrating the access of the various input interfaces to the pool of VCIs available at each output port, thus allowing the implementation of a fully distributed control architecture. The multiplex switch described in [51] can be considered at the top level as a 3-stage Clos arrangement, with module sizes 4 x 4, 64 x 64 and 4 X 4 respectively. The inner 64 x 64 modules are realized with parallel planes of banyan networks made of 4 x 4 elements: plane parallelism and an internal 4:l speed advantage make these modules nonblocking.…”
Section: Nttmentioning
confidence: 99%
“…In the VLSI implementation described in [51] a 0.8 pm BiCMOS process is employed, which allows a speed of up to 620 Mbit/s on every 4 bit parallel port. The buffer size is only 31 cells.…”
Section: Nttmentioning
confidence: 99%