Multiplex communications – Communication techniques for information carried in plural... – Combining or distributing information via time channels
Reexamination Certificate
1999-04-23
2002-12-24
Olms, Douglas (Department: 2661)
Multiplex communications
Communication techniques for information carried in plural...
Combining or distributing information via time channels
C370S411000
Reexamination Certificate
active
06498798
ABSTRACT:
FIELD IF THE INVENTION
The present invention relates to data communications systems.
BACKGROUND INFORMATION
Because of the changing nature of Local Area Network (LAN) traffic due to client-server and web-based applications, larger fractions of data are exiting individual workgroups for destinations outside the workgroups. With conventional approaches, a hub (or repeater) operating in accordance with a Carrier-Sense Multiple Access/Collision Detect (CSMA/CD) protocol is typically used to provide arbitration for the shared interconnection network.
Conventional Ethernet hubs or repeaters rely on CSMA/CD protocols being implemented on all incoming and outgoing ports, i.e., a conventional repeater or hub simply broadcasts the input arriving on a channel to all ports which are then able to apply the CSMA/CD protocol arbitration to determine when they are able to access the shared 10 Mbit/s (in the case of Ethernet) or 100 Mbit/s (in the case of Fast Ethernet) channel in half-duplex mode.
The connectivity paradigm in LANs is rapidly changing due to the surge in world-wide-web based traffic (in corporate Intranets and the Internet) and other client-server based computing applications. The old 80/20 rule-of-thumb for data traffic, where approximately 80% of data traffic stayed within the workgroup and only 20% of the traffic traveled the backbone, is quickly reversing to a 20/80 paradigm where most of the data traffic leaves the individual workgroups.
At higher data rates, a potential alternative to CSMA/CD-based links are dedicated point-to-point buffered full-duplex links. Such an approach may remove the need for conventional half-duplex repeaters. One type of device is a full-duplex buffered repeater such as the Gigabit Ethernet Buffered Repeater offered by Packet Engines. This device takes full-duplex ports as its inputs and performs CSMA/CD arbitration within the confines of a box that is only 1-2 meters wide, thereby allowing Gigabit speeds over lengths of hundreds of meters. Furthermore, this device allows several lower speed ports (e.g., 10-100 Mbit/s) to share a higher speed (e.g., 1 Gbit/s) output port.
Nonetheless, approaches that use the CSMA/CD protocol, such as the aforementioned full-duplex repeater, are subject to efficiency limitations. For instance, as the network load of a typical CSMA/CD link increases (e.g., to 50% or higher), the efficiency of the link is drastically impaired because most of the ports are sending colliding messages a large fraction of the time. As this happens, no single port can take control of the channel, and channel efficiency drops.
An alternative approach is the use of a multiplexer-hub, as shown in
FIG. 1. A
conventional multiplexer
110
, as shown in
FIG. 1
, simply samples data consecutively from a plurality of lower-speed input lines and places them bit (or byte) interleaved onto a high-speed output line. A limitation of this approach, however, is that it may not always make the most efficient use of the bandwidth of the multiplexed link in real data networking situations, where the traffic is bursty (i.e., data packets arrive in bursts) and arrives at random intervals. The simple conventional multiplexer
110
does not provide any arbitration for access to its output link, so the output link must be provisioned with a capacity that is much greater than the input links. For instance, multiplexing data from N input lines onto one output line usually requires that the output line operate at N times the bit-rate of any of the input lines. Because it is not possible to determine a priori which input lines contain “live” data and which input lines are not carrying useful (or live) data—thanks to the random nature of packet arrivals at the inputs—it is quite possible that the live data is multiplexed onto the links along with data from the idle links that are not carrying live packets at that particular moment. Because arbitration is lacking, the high-bandwidth (and high-cost) link resource is hence not used optimally, thereby limiting the ability of many users to share the resource.
A simple multiplexed link, such as that of
FIG. 1
, while removing the inefficiency of the conventional CSMA/CD arbitration (at high loads) for aggregated traffic, may not necessarily represent a performance/cost improvement over the conventional CSMA/CD based hub discussed above.
SUMMARY OF THE INVENTION
The present invention provides a scaleable multiplex link which overcomes the limitations of conventional CSMA/CD and multiplexed systems. In an exemplary embodiment of a hub in accordance with the present invention, a statistical multiplexer, or concentrator, is used to efficiently combine multiple lower-speed data streams into a smaller number of higher-speed data streams. The higher-speed data streams are further combined in a multiplexer into one high-speed data stream.
The multiplexer-hub of the present invention is better suited to current trends in data traffic where most of the traffic within a sub-network exits due to an increase in client-server type traffic. A multiplexed link in accordance with the present invention is easier to implement while being more efficient than conventional arrangements.
REFERENCES:
patent: 5166926 (1992-11-01), Cisneros et al.
patent: 5461626 (1995-10-01), Takase et al.
patent: 5541914 (1996-07-01), Krishnamoorthy et al.
patent: 5768270 (1998-06-01), Ha-Duong
patent: 5875190 (1999-02-01), Law
patent: 5949778 (1999-09-01), Abu-Amara et al.
patent: 6026092 (2000-02-01), Abu-Amara et al.
IEE Proc.-commun., vol. 145, No. 3, Jun. 1998—Design of knockout concentrators by Y.-S. Lin et al.*
K.S. Chan et al., “Clos-Knockout: A large-scale Modular Multicast ATM Switch”, 1996 IEEE.*
Minze V. Chien et al., “Adaptive Binary Sorting Schemes and Associtated Interconnection Networks”, IEEE transactions on parallel and distributed systems, vol. 5, No. 6, Jun. 1994.*
Hyunsoo Yoon et al., “The knockout switch under nonuniform traffic”, 1988 IEEE.*
IEEE Journal of Solid-State Circuits, vol. 25, No. 6 Dec. 1990—A CMOS Batcher and Banyan Chip Set for B-ISDN Packet Switching by William S. Marcus.
IEEE Journal On Selected Areas in Communications, vol. SAC-5 No. 8 Oct. 1987—The Knockout Switch: A Simple, Modular Architecture for High-Performance Packet Switching by Yu-Shuan Yeh, fellow, IEEE, Michael G. Hluchyj et al.
Krishnamoorthy Ashok V.
Nuss Martin C.
Baker & McKenzie
Olms Douglas
Sam Phirin
LandOfFree
Priority-based statistical multiplexer-hub does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Priority-based statistical multiplexer-hub, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Priority-based statistical multiplexer-hub will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2922190