Architecture for lightweight signaling in ATM networks

Details Architecture for lightweight signaling in ATM networks Architecture for lightweight signaling in ATM networks

2000-10-24

2001-06-05

Patel, Ajit (Department: 2662)

Multiplex communications

Pathfinding or routing

Switching a message which includes an address header

Reexamination Certificate

active

06243387

ABSTRACT:

BACKGROUND
The present invention is directed to a method and apparatus for providing a new signaling architecture for a communication network. More specifically the present invention provides for lightweight and extensible signaling in Asynchronous Transfer Mode (ATM) networks.
Modern communication networks need to support a wide range of applications with diverse service requirements. Network technologies that were traditionally designed and optimized for specific communication services are evolving towards accommodating this diversity on a single integrated communication infrastructure. Both Internet Protocol (IP) and Asynchronous Transfer Mode (ATM) networks have been moving in the direction of supporting a broad range of multimedia flows, possibly from different starting points. Both of these architectures support Quality of Service (Q of S) and integral to their approach is a notion of identifying flows, albeit with different amounts of persistent state being maintained in the network.
Connection oriented networks have been developed over the years for supporting applications that require stringent quality of service, and the ability to exploit switching to achieve good price/performance characteristics for inter-networking. Switching has continued to maintain this superior price/performance in spite of the dramatic changes in system and semiconductor technology over the past decade. Several ongoing efforts to utilize switching for IP interconnection are grounded in the belief that these superior price/performance characteristics will continue to hold. The ability to use a short handle (e.g., a virtual circuit identifier (VC id)) to forward data is based on the desire to route once (a relatively expensive operation) and then switch (a cheaper operation) many times. ATM is a connection-oriented technology that has been designed to use switching for “scalable bandwidth”. It has been proposed as an attractive technology for flexible support of guaranteed quality-of-service. At the same time, the scalability of a design with perflow state has been questioned often, with good justification. It is clear that with tremendous number of flows impinging at some point, especially at the core of the network, flow aggregation is mandatory. Virtual paths in ATM networks offer one solution, see The ATM Forum, “ATM User Network Interface (UNI) Specification, Version 3.1, Prentice Hall 1995. The aggregation techniques espoused for integrated services in the next generation IP networks are another, see “Link-Sharing and Resource Management Models for Packet Networks, ” S. Floyd et al., IEEE/ACM Transactions on Networking, Vol. 3, No. 4, pp. 365-386 August 1995. These solutions favor avoiding signaling altogether at the cost of not being able to respond to the special needs of an individual flow. While this may be acceptable in some cases, there remains the need to improve signaling for a significant set of flows.
An example of a conventional ATM signaling procedure is illustrated in FIG.
1
. In this configuration a calling station
101
seeks to setup a call to called station
110
. In this setup operation the call is connected between the stations by intermediate switches
103
and
105
. The part of the connection between the calling station
101
and switch
103
is referred to as a first hop. The part of the connection between the switches
103
and
105
is referred to as a second hop. The part of the connection between switch
105
and the called station
110
is referred to as a third hop.
In operation a complex setup message that includes, not only that which is necessary for fundamental connectivity, but also the requirements for Q of S, and other user-related functions, is sent by the calling station
101
to switch
103
at time t
l
. The switch then must process the setup message, including performing admission control with reference to the Q of S requirements, before it forwards the setup message to switch
105
. Similarly, switch
105
must process the setup message before it forwards that message to the called station. Involved in this processing is a determination, at the first switch, of the end to end route. This decision process is very complex because of the inclusion of “reassembly”, route determination, and Q of S processing all at once.
The setup message is considered the repository of information to be communicated from the end-systems to the network for any and all additional functionality desired from the network for the call. This not only leads to a relatively large setup message, but the need to incorporate substantial facilities or information in the message structure to accommodate known and even potentially unforeseen requirements.
The switch, before routing the setup message, must reassemble the setup message and perform the potentially complex task of determining the Q of S requirements for the call being established. This information may potentially be used to both determine a Q of S compatible route and perform admission control functions. In addition, when a setup message arrives at the switch, the switch has to send a message back to the upstream node indicating a “call proceeding” state, so as to enable the upstream node to maintain short timers.
It is only when the setup message flows all the way to the destination, the called station, and a “Connect” message is returned all the way back from the called station to the calling station that the actual data flow may be started. Additionally an acknowledge message, a “Connect-Ack” message, has to flow from the upstream node (switch or end-system) to the downstream node to indicate to the receiving end that data may now begin to flow.
With this traditional connection setup, the use of the path is delayed until the connection setup is complete from end to end. One of the disadvantages is that this reduces resource utilization as the resources are allocated and held before they can be used. Studies of reservation schemes and connection setup on the Internet indicate that this penalty can be significant. Moreover, the Private Network to Network Interface (PNNI) signaling proposal for ATM may not be suitable for global scale networking, and yields low utilization of network resources unless precise traffic characteristics are known. The challenge in building a single network for both connectionless and connection-oriented services is to provide a protocol that nullifies complexity and the delay penalty of call setup, while supporting service guarantees and intra-domain manageability.
ATM's signaling mechanisms are woefully inadequate for the type of short transactions that typify current-day data communications, such as on the Internet. An interaction between a source and destination may comprise only a few packets of data. A majority of the flows in the Internet fit this pattern. The overhead of a heavy-weight connection setup is only amortized when there is a sustained interaction involving the exchange of a large number of packets. But this is true only for a relatively small number of flows. One approach is to use an existing virtual circuit (VC) to multiplex flows between a pair of routers, with the hope of aggregating multiple short transactions between different systems. This VC is used to carry flows that have similar characteristics. This potentially leads to a N*N virtual interconnection among multiplexing points, and when flows desire to be treated differently, it may lead to an even larger set of logical connections maintained on a semi-static basis.
Signaling also has been viewed as being too complex, with multiple messages required to setup a connection. Also, the complex encoding of signaling messages requires considerable processing. The desire to provide sufficient flexibility in the signaling process as may be required by any and all applications that are likely to use the network is one cause for the introduction of this complexity. Furthermore, current design of ATM or Signaling System 7 (SS7) signaling requires hop-by-hop processing of the setup message through the entire end-to-en

Affiliated with

Also associated with

Rating

Architecture for lightweight signaling in ATM networks does not yet have a rating. At this time, there are no reviews or comments for this patent.

If you have personal experience with Architecture for lightweight signaling in ATM networks, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Architecture for lightweight signaling in ATM networks will most certainly appreciate the feedback.

Rate now

Comments { 0 }

Profile ID: LFUS-PAI-O-2447281