Multiplex communications – Communication techniques for information carried in plural... – Transmission bandwidth conservation
Reexamination Certificate
2000-03-03
2004-04-06
Pham, Chi (Department: 2663)
Multiplex communications
Communication techniques for information carried in plural...
Transmission bandwidth conservation
C370S389000
Reexamination Certificate
active
06717956
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to communication networks and, in particular, to an addressing scheme for a network to reduce overhead.
BACKGROUND
Data-carrying capacity in access and long-haul networks is a billable commodity to service providers. Traditional networks have employed a single static addressing mode for data link layer and network layer devices in these networks, such as 32-bit Internet Protocol (IP) addresses or 48-bit Media Access Control (MAC) addresses in Gigabit Ethernet. The motivation for long addresses is that every device across all networks worldwide can be assigned a unique data link layer and/or network layer address, which enables full portability of devices without address duplication conflicts and with a minimum of management overhead. However, these addresses compose a large portion of the packet header overhead (roughly 60% for Gigabit Ethernet and roughly 40% for IP) added to packets at the networking layer or at the data link layer. Any reduction in this overhead through compression of these addresses increases the data-carrying capacity of such networks.
In integrated voice and data networks, the average size of IP data packets is roughly 250 bytes, with over 50% of the packets being only 64 bytes. The average size of circuit-emulated voice packets is usually smaller than the average size of IP packets to minimize packetization delay (assume 150 bytes). The 12 bytes of MAC-layer addressing is a significant fraction (4% of data, 7% of voice) of the overall packet size, and thus any compression of this addressing will significantly improve the data-carrying capacity of deployed access and long-haul networks using an Ethernet-like MAC layer. This directly adds to the billable capacity of the service providers that own such networks.
Related to the invention described herein of dual-mode addressing is the identification of the network topology. Topology reconfiguration scenarios include network initialization, insertion of devices, deletion of devices, topology changes that do not involve insertion or deletion of devices (such as link breaks, where a link connects a pair of devices), and combining of operating networks. In the context of this document, network initialization does not refer to the internal processes independently used by each device to initialize itself, but rather to the communication between interconnected devices required to establish knowledge of network topology and remapping of short addresses to long addresses.
There are several fundamental requirements that must be met by the mechanisms used for topology reconfiguration:
1. Ongoing traffic between unperturbed devices on networks undergoing reconfiguration shall continue to flow, assuming that there are multiple paths available for such traffic. In the event that the reconfiguration involves the temporary removal of physical routes on which traffic was flowing, standard protection switching mechanisms such as SONET-based line or path switching or other mechanisms for packet-switched networks are used to temporarily reroute the traffic to unaffected physical routes between nodes.
2. The mechanism shall be plug-and-play, e.g. determination of topology changes shall occur automatically and shall not require intervention from network management systems.
3. The communication mechanism between devices shall enable topology change information detected by a given device to propagate to all other devices on the virtual network. This can be done using a standard topology discovery mechanism or using other mechanisms. The choice of mechanism is based on the specific requirements of the individual virtual network.
Many current topology discovery mechanisms are distributed in the sense that each device in the network constructs and stores its own version of the network topology based on information received from other devices about their own neighboring devices, referred to in this document as neighbor status messages. A good example of such a mechanism is the link state protocol for broadcast of topology changes used in the OSPF routing protocol, described in the book “Interconnections, Second Edition” by Radia Perlman, Addison Wesley Longman, Inc., 2000, incorporated herein by reference in its entirety. The link state protocol, along with all other distributed topology discovery protocols known to the authors, relies on the mechanisms of age-out of topology information and reliable delivery using acknowledgement messages sent from the device receiving a topology message back to the source of the message.
The use of an age-out mechanism means that the topology stored at any device will become invalid after a configurable period of time. This means that neighbor status messages must be periodically sent by every device in the network, even if there is no change in the topology. This is inefficient both in terms of processing at each device and in terms of network bandwidth because changes in network topology are not frequent occurrences. A mechanism that removes the necessity for each device to age-out its topology would therefore be useful.
The use of reliable delivery of neighbor status messages through tracking of received acknowledgement messages at each device is a standard approach to ensure that all transmitted neighbor status messages are received, and thus that all devices construct a correct network topology. There remain transient scenarios, however, such as devices going down and coming back up, that can result in some devices not receiving all messages, and thus constructing an incorrect network topology that can result in other devices on the network becoming invisible. In packet-switched data networks, there has traditionally been no guarantee of reliable service, and thus no additional mechanisms to guarantee construction of a valid network topology at each device have been required. To transport telco-quality voice on DS
1
or DS
3
leased lines over a packet-switched network, however, extremely high reliability is required. A mechanism that validates the topology constructed at each device would therefore be useful.
There are currently no established mechanisms for topology reconfiguration in networks using dual mode addressing. What is needed for this type of networks is a mechanism that:
1. Enables topology reconfiguration and that meets the above general topology reconfiguration requirements.
2. Minimizes (and preferably eliminates) changes to management and control information (such as provisioning tables and routing tables internal to a device) due to switching between dual addressing modes necessitated by reconfiguration.
3. Enables re-establishment of short addresses as part of reconfiguration, e.g. that ensures the elimination of short address duplication when multiple networks are combined together.
SUMMARY
A dual addressing mode is described in which reduced-length addresses (referred to as short addresses) are substituted for standard addresses (referred to as long addresses) for traffic whose source or destination is internal to a given virtual network topology. Long addresses are globally unique, while short addresses are locally unique to a given virtual network. The required length of short addresses used for a given virtual topology is dependent on the number of devices reachable within the topology. For a virtual topology with less than 256 addressable devices, for example, 8-bit short addresses can be used.
When a node within the virtual network sees a packet with a short destination address in the header, the node understands the address to be within the virtual network and routes the packet accordingly. A node is defined as a point where traffic can enter or exit the ring. If a source address is a short address, the virtual network can identify the source within the virtual network. For packets originating in the virtual network whose destination is also in the virtual network, both the source and destination addresses can be short addresses.
A node within the virtual network that is also connected to an outside network may, optionall
Bannai Vinay K.
Fan Jason C.
Jogalekar Prasad P.
George Keith M.
Luminous Networks, Inc.
Pham Chi
Sawyer Law Group LLP
LandOfFree
Dual-mode virtual network addressing does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Dual-mode virtual network addressing, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Dual-mode virtual network addressing will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3242920