Electrical computers and digital processing systems: multicomput – Multicomputer data transferring via shared memory – Partitioned shared memory
Reexamination Certificate
1999-05-14
2001-11-27
Dinh, Dung C. (Department: 2153)
Electrical computers and digital processing systems: multicomput
Multicomputer data transferring via shared memory
Partitioned shared memory
C709S201000, C455S433000, C455S435100, C714S011000
Reexamination Certificate
active
06324572
ABSTRACT:
RELATED FIELD OF THE INVENTION
The present invention generally relates to communication systems, particularly, a network communication system.
BACKGROUND OF THE INVENTION
In a network communication system, failure of a service connection is a type of problem that in particular is associated with failure of a server or its interconnection. Failure of a service connection is highly undesirable in many types of service connections. To reconnect a failed service connection, the client of the service connection requests from a home server to re-build the service connection that may take up to 16 seconds. In case of data service connection, 1-16 seconds laps of time for re-building the connection service may not be a problem if the data reception is immune to a latency deficiency. However, in case of voice or video service connection, transportation latency is a major problem because communication of video or voice requires very low latency to maintain a coherent voice or video communication between two end users of the service connection. To overcome this problem, a fault tolerant network system with redundancy is implemented at high cost and complexity.
Referring to
FIG. 1
, a conceptual block diagram of a communication network according to a prior art is shown. To make a service connection, one of the clients C
1
-N,
110
-
1
through N, for example
110
-
1
, requests through an interconnection
130
-
1
for service connection from a home server
101
. Home server
101
has a general connection
150
with a distributed network
190
that includes a number of servers S
1
-M,
120
-
1
through M. Home server
101
includes network addresses of all servers in the distributed network
190
. After one or more exchanges of information associated with the service connection, a service connection
140
-
1
may be assigned with at least one of the servers
120
-
1
through M, for example a server
120
-
1
, to provide the requested service connection. The server
120
-
1
normally is identified by a network address. Subsequently, client
110
-
1
communicates directly through service connection
140
-
1
with server
120
-
1
for receiving the services it may require. Clients
110
-
1
through N may include any computing device such as a computer, or communication devices such as a telephone, or an interactive display device such as a television, or a combination of computing, communication and display devices. The service may include service connection for transporting data, video or voice information.
If a service connection, such as service connection
140
-
1
fails for any reason, the client
110
-
1
communicates with the home server
101
to receive an address of another server, selected from the servers
120
-
1
through M, to continue the service. The delay in rebuilding a failed service connection in case of voice and live video connection services is highly undesirable. The delay purposely reduces quality of voice and video service connections. In case of data service connections, when data latency is a factor of quality of service connection, the delay is again highly undesirable.
According to a prior art, a way of improving the fault tolerant of the communication network system is to assign two servers to a communication service connection such that one server functions as the active server and another as a standby server in a synchronous manner. If the active server fails for any reason, the standby server, while being in synchronous with the active server, would take over the service connection activities and provides possibly an adequate and minimally uninterrupted service connection. The standby server in this case has an independent central processing unit (CPU) that works in parallel and synchronous with a CPU of the active server. The standby CPU copies and follows in most part or completely all activities of the active CPU. In case of a failure of the active CPU that causes failure of the service connection, the standby CPU takes over the service connection activities and allows a minimally uninterrupted service connection. The interruption may last for a range of 1-3 mSec. This method may involve, nevertheless, the client and home server interaction to successfully transfer the processing of the service connection from the active to the standby CPU. Keeping a standby CPU active and in synchronous along with the active CPU adds complexity and cost to the network communication system.
Referring to
FIG. 2
, to build a communication network system with relatively higher fault tolerance according to another prior art, a number of servers may be clustered together to provide the connection services. Such an embodiment is shown by way of incorporating a common cluster block
210
which provides an addressing and switching scheme as to make more than one server in the cluster available for service connection at a time. For example, a service connection
240
-
1
passes through common cluster
210
to reach a server in the distributed network
190
, shown as a cluster in this case, through connection
220
. If an initially assigned server fails to provide the service connection, another server selected from a common cluster takes over the service connection with minimal interruption. The change in the server in this case would in most part or completely is transparent to the client
110
-
1
that originated the request or receiving the service connection
240
-
1
. The common cluster
210
assigns a new address of the selected server to the service connection
240
-
1
. The client
110
-
1
communicates with the common cluster
210
which would be at a fixed address regardless of the server selected for servicing the service connection
240
-
1
. As a result, the failure of the server becomes transparent to the client
110
-
1
which eliminates a need to contact the home server
101
for rebuilding the service connection. In turn, a substantial reduction in the latency of reconnection is achieved for providing a continuous service connection when a service connection fails. As such, such common cluster redundancy improves the fault tolerant of the communication network. However, such redundancy and common clustering of the servers is not without cost and complexity. A fault tolerant network communication system is implemented at very high cost and complexity.
Recently, traditional communication systems that mainly used hierarchical network structures have moved to a more distributed structure such as an internet structure to communicate voice, data and possibly video over such distributed structures. Traditional communication networks, particularly in cellular application, included a number of base station transceiver sites (BTS), a number of base station controllers (BSC), and a number of mobile station controllers (MSC). A mobile station initiating a wireless connection with a BTS is forced to utilize BSC and MSC resources that have rigidly provided support services to that BTS. As a result, when resources of a BSC reduce to low levels, the mobile station must change the BTS in order to receive resources from another BSC which may force an interruption of service connection to the mobile station. Moreover, such hierarchical structures in some cases may not provide an optimum usage of available resources. A distributed network may nevertheless be very expensive to provide high fault tolerance network for voice communication in a cellular application which requires low latency.
Therefore, there is a need for a communication network that provides high fault tolerance performance at a minimal cost.
REFERENCES:
patent: 5774660 (1998-06-01), Brendel et al.
patent: 5774668 (1998-06-01), Choquier et al.
patent: 6163866 (2000-12-01), Shrivastava
Liss Raymond M.
Silverman Shmuel
Stewart Randall R.
Beladi Sayed Hossain
Dinh Dung C.
Motorola Inc.
LandOfFree
Communication network method and apparatus does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Communication network method and apparatus, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Communication network method and apparatus will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2592946