Multiplex communications – Communication techniques for information carried in plural... – Combining or distributing information via time channels
Reexamination Certificate
1997-12-31
2001-02-27
Rao, Scema S. (Department: 2662)
Multiplex communications
Communication techniques for information carried in plural...
Combining or distributing information via time channels
C370S420000, C370S463000, C370S907000, C359S199200, C359S199200
Reexamination Certificate
active
06195367
ABSTRACT:
FIELD OF THE INVENTION
This invention pertains generally to the field of bandwidth management in high-capacity communication networks, and particularly to the problem of managing the large amounts of bandwidth flowing through the central offices of such networks.
BACKGROUND OF THE INVENTION
Vital to any telecommunication network is the manner in which its busiest points manage great quantities of bandwidth. These points usually occur at large central offices (COs). At COs, enormous amounts of traffic must be switched between enormous numbers of possible input or output links of varying data capacities.
Space-switches are vital to achieving this function, since they interconnect the large numbers of input and output links that converge at a CO. Presently, these space-switches are typically implemented using digital cross-connect (DCS) technology. DCS switches switch signals that are in electrical form and that conform with one of a set of digital signal formatting protocols. The highest capacity signal that can be switched by a DCS switch complies with the DS3 protocol, which supports a data rate of 44.736 Mbps. DCS switches capable of supporting this maximum rate are referred to as DCS 3/3 switches, and are widely used in present large COs.
With the emergence of optical fibres as the principle transport medium for carrying telecommunication signals from point to point however, the exclusive use of DCS space-switches is becoming an undesirable means by which to space-switch large quantities of traffic within very busy COs. This is firstly because the signals sent through fibre optic media are in optical form rather than in electrical form, which means that the use of DCS switches to guide a signal from an optical input link to an optical output link mandates the conversion of many of the signals from an optical form to an electrical form for switching purposes, and then back to an optical form for transport purposes. Such conversions waste resources and add complexity to the design of the CO and its components.
A second and more important reason why the exclusive use of DCS technology is no longer suitable for deployment in busy COs, is because it has not yet been cheaply adapted to switch individual signals that are operating at rates higher than the rate supported by the DS3 protocol (44.736 Mbps). This is a problem because all optical signals comply with protocols such as the Synchronous optical Network (SONET) protocols, that call for rates many times greater than 44.736 Mbps. The limit on the per-port throughput of DCS switches means that the only way to use them to space-switch a signal between two links supporting rates greater than 44.736 Mbps is by demultiplexing the higher-rate signal down to the DS3 rate for switching purposes, switching it on existing DCS 3/3 switches at the DS3 rate, and then remultiplexing the signal back up to a higher-rate for transport purposes. Like the first problem, this mandated practice also leads to a waste of resources and increased complexity in the design of the CO and its components. Moreover, this limitation in the per-port rates supported by DCS switches means that each fibre-optics link requires tens or hundreds of cables to interface with the highest capacity DCS switch, which results in an enormous number of cables having to be laid in each CO.
Even more significantly, this limitation in the per-port rates supported by DCS switches places limits on the total bandwidth that can be managed by a CO, which relies only on DCS switches to interconnect various pieces of network elements More specifically, this limitation means that when a signal operating at a rate higher than DS3 needs to be space-switched across a DCS 3/3 switch, it must be demultiplexed into several lower-rate signals each operating at the DS3 rate. Each of these lower-rate signals require an input port and output port on the DS3-X switch. This means that many ports could be needed to connect a single high-rate link to other network elements within the CO. For example, 48 input and output DCS 3/3 ports are required to connect a single OC-48 link to other parts of the CO. Support for a plurality of such high-rate links can thus quickly lead to the utilization of all ports on an existing DS3-X space switch, which comes in sizes larger than 1000 ports by 1000 ports at a great cost, and which does not come in sizes larger than 2000 ports by 2000 ports. Once all the ports on a space-switch are utilized, no further interconnections between links that converge at a CO are possible. This in turn imposes a limit on the amount of bandwidth that can flow through the CO.
Cascading DCS 3/3 switches to form single higher-capacity space switches does not significantly alleviate this last problem, as no further increases in switching capacity are realized once more than a few switches have been interconnected. More specifically, as the number of switches comprising a single larger block of switches increases, each additional switch that is added to the block must set aside an increasing proportion of its ports just to communicate with the other switches in the block, rather than to connect new network links into the CO. At some point, the addition of an additional switch would fail to add any switching capacity to the CO, as the additional switch would have to allocate all of its ports just to communicate with the other space-switches that are already in the block. Therefore, even if deployed as interconnected groups, the use of DCS switches impose a limit on the total bandwidth that can be handled by a CO.
A solution to the problems created by the use of DCS switches is thus required. The use of electrically controlled optical signal cross-connect (OXC) switches, in which an output fibre carries the same light that entered the switch on an fibre, could eventually comprise a complete solution to the problem of managing large amounts of bandwidth at busy nodes. By switching signals in their optical forms, these switches avoid the signal conversion problems associated with DCS technology. Furthermore, because such a switch is capable of interconnecting links operating at any of the high-capacity SONET rates, it can also increase the bandwidth capacity of the CO.
At present however, problems with cross-talk, the lack of a memory function when processing optical signals, and a lack of reliability, are restricting the size of these OXC switches to 16 ports by 16 ports, and preventing them from being interconnected to form a larger switch. These restrictions are becoming unacceptable for space-switches that are to be deployed in very busy COs at which several tens of high-capacity links need to be interconnected.
Additionally, there are no known methods for performing several important bandwidth management services on optical signals, such as grooming, time-slot interchanging and performance monitoring, that are routinely performed on electrical signals. Therefore, an OXC switch by itself is not a desirable solution to the aforementioned bandwidth management problems.
Another proposed solution to these problems, involves providing optical interfaces into a DCS switch. This would internalize much of the signal processing required to electrically switch signals that arrive at and leave from the CO in optical form, within the DCS switch. It also allows for a reduction in the number of cables required to connect network elements to the DCS switch. Moreover, since the DCS switch would still switch electrical signals, the CO would still be able to support a variety of bandwidth management services.
At present however, the highest available interfaces support only the OC-12 rate. DCS switches with OC-12 interfaces are also very costly. Furthermore, this approach does not by itself address the previously described total bandwidth limitation associated with DCS technology. This is because even when DCS switches are fitted with optical interfaces, they still switch signals in a low-rate electrical format that cannot support rates exceeding 44.736 Mbps. Therefore, even if OC-12 interfaces are provi
Barsetti Sarto
Betts Malcolm
Gibbon Mark Acton
Jakobik Bogdan
Liu Xinyi
Nortel Networks Limited
Rao Scema S.
Tsegaye Saba
LandOfFree
Architectural arrangement for bandwidth management in large... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Architectural arrangement for bandwidth management in large..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Architectural arrangement for bandwidth management in large... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2560469