Multiplex communications – Fault recovery – Bypass an inoperative switch or inoperative element of a...
Reexamination Certificate
2000-07-17
2004-10-19
Duong, Frank (Department: 2666)
Multiplex communications
Fault recovery
Bypass an inoperative switch or inoperative element of a...
C370S221000, C370S225000, C370S395530, C370S401000, C370S463000, C370S465000
Reexamination Certificate
active
06807149
ABSTRACT:
TECHNICAL FIELD
The present invention relates in general to data processing systems, and in particular, to Asynchronous Transfer Mode (ATM) networks using Local Area Network (LAN) emulation and LAN Emulation Clients (LECs).
BACKGROUND INFORMATION
Asynchronous Transfer Mode (ATM), as a networking technology, has been gaining in popularity. It is widely perceived to be the underlying technology for high speed networks of the future. ATM's strengths are that it is highly scalable, in terms of access speeds (from 1.5 Mbps to as high as 1.2 Gbps and more), and in terms of application traffic (voice, video, and data). It also has its drawbacks, the primary one being its complexity and overheads, from the physical level (e.g., when Synchronous Optical Network (SONET) framing is used) to the ATM level (e.g., traffic and congestion control) to higher level functions (e.g., signaling).
As with any technology, a big challenge is to find applications where one can justify the cost of moving to the new technology. In many cases, network managers are hesitant to change to a new technology. If there is not enough demand for a new technology, equipment costs will remain high enough to prevent entry. With ATM there are many expected important applications; most of them are variations of high bandwidth and multimedia applications (e.g., scientific visualization, distance learning, video on demand, multimedia conferencing, etc.). These can be termed “native mode” applications as they will communicate directly over an ATM network. However, the widespread availability of such applications will not occur unless ATM networks become more ubiquitous (so end-to-end ATM connections can be established) and equipment costs decline sufficiently so that consumers can afford to deploy the important applications.
In the meantime, there are fewer glamorous applications using ATM today. These applications exist on today's networks that can utilize the high bandwidth inherent in ATM networks. For existing applications to work over ATM networks, an “emulation” layer has to be inserted which emulates the services of existing networks over ATM networks such that existing applications continue to operate as if they were using a higher speed version of old networks. The benefit of using ATM in this scenario is, existing applications receive high speed access and they may continue to interwork with existing networks.
To address the issue stated above, the ATM Forum (a committee to address ATM implementation issues) developed LAN Emulation which purports to run an emulation layer atop ATM hardware which makes the ATM network look like a conventional LAN, for example an Ethernet or a Token Ring network. With LAN Emulation, users can continue to run existing LAN applications while migrating their hardware to the higher bandwidth supported by ATM. LAN Emulation provides an easy migration path for network managers to move toward adopting ATM technology.
To define a LAN Emulation service over ATM, one has to consider the differences between ATM networks and other networks (e.g., IEEE 802.3 and 802.5 networks). Some of the major differences are listed below:
LANs offer connectionless service, so applications can communicate without a prior setup. ATM networks are connection oriented; before two end systems can communicate, a connection must be established via signaling (SVC) or via provisioning (PVC).
LANs are broadcast media, and ATM networks are point to point.
LANs use a flat 48 bit address space. ATM networks use larger more hierarchical addresses.
LAN boundaries, physical and administrative, are typically the same.
In contrast, ATM networks are point to point and as such arbitrary administrative domains may be defined independent of physical topologies.
Because of the above differences, the LAN Emulation layer must provide a connectionless data transfer, simple broadcast/multicast, address mapping of network addresses, and the ability to interconnect with existing LANs.
LAN Emulation is based on a client-server architecture, hence there are two types of components:
LAN Emulation Clients (LEC); and
service components LAN Emulation Servers (LES), Broadcast and Unknown Server (BUS) and LAN Emulation Configuration Server (LECS).
Any ATM end system that wishes to use the services of LANs must have a LEC for a particular type of LAN. Such ATM end systems are of two types.
ATM Hosts: End systems that contain the applications that wish to communicate with other (LAN or ATM) end systems.
Edge Devices: These are bridges or routers that forward data from the emulated LAN (ELAN) to other LANs (emulated or otherwise).
An ELAN consisting entirely of ATM hosts, allows communication between applications on ATM attached devices only. An ELAN consisting entirely of “Edge Devices”, allows LANs (e.g., “legacy” IEEE LANs) to be connected to each other across an ATM network. An ELAN that has a mix of ATM hosts and Edge Devices allows communication between ATM attached devices and legacy LAN stations.
A LEC communicates with other LECs to transfer data between communicating applications. To identify a destination for a data transfer, a LEC uses LAN emulation service components to translate a matching network (MAC) address to an ATM address.
Data communication using LAN Emulation to communicate on ATM networks is quite complex as is apparent in the above discussion. While the usefulness of ATM is apparent, the complexity of the network connections requires methods for ensuring reliability of connections and methods for “failover” (set up alternate communication paths in a soft fail mode) in the case a network connection fails. The failover performance determines the “resiliency” of the connection.
When network connections are established it may be possible to create two or more communication paths linking one segment to another. This creates a loop where two bridges on two LANs pass the same data from LAN to LAN forever. To prevent more than one communication path from occurring, transparent bridges use a spanning tree algorithm or protocol to determine a unique path between a source and a destination, avoiding bridge loops. In Edge Devices for enterprise ATM networks, LEC resiliency for example, is typically achieved by defining two LECs on different ATM adapters. These LECs both join the same ELAN and both are added as ports on the Edge Device bridge. A spanning tree protocol would block one of the two LEC ports to prevent a loop from occurring in the bridged network. If the ATM connection of the active LEC fails, the normal spanning tree procedures will activate the blocked LEC. In this case, the failover time of the spanning tree protocol can be quite long, typically on the order of forty-five seconds. This time is long enough to cause the failure of active sessions. There is a clear need for a way to provide improved LEC failover while keeping the response time low.
SUMMARY OF THE INVENTION
A LEC (any ATM end system that wants to use the services of IEEE 802.3 or 802.5 LANs) would have a connection typically defined through a bridge (Edge Device) to the device to which it is communicating. To insure connection resiliency, two LECs are defined in normal practice with a lengthy and extensive recovery procedure enforced in case of an ATM interface failure. In the present invention, only one LEC is configured. This LEC is configured to be associated with a given ATM interface with an alternate ATM interface also being defined. If the active ATM interface fails, the bridge is not notified of the failure as would normally happen. Bridge notification and the lengthy spanning tree topology change procedures are avoided. Instead, the LEC associates itself with the alternate ATM and rejoins the ELAN via the alternate ATM interface. Typically the ELAN Join Procedure can be completed in less than one second. The ELAN Join Procedure is one of the LAN Emulation protocols. Using this technique, failover may be performed with a sub-second response time. There is no need to inform the bridge that the failover had occurred since there was no cha
Alexander, Jr. Cedell Adam
Karlsson Keith Edward
Cockburn Joscelyn G.
Duong Frank
International Business Machines - Corporation
Winstead Sechrest & Minick PC
LandOfFree
Method and system for LEC resiliency with fast failover does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and system for LEC resiliency with fast failover, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and system for LEC resiliency with fast failover will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3309903