Multiplex communications – Diagnostic testing – Determination of communication parameters
Reexamination Certificate
1997-11-04
2001-03-27
Vu, Huy D. (Department: 2735)
Multiplex communications
Diagnostic testing
Determination of communication parameters
C370S235000, C370S248000, C370S351000
Reexamination Certificate
active
06208622
ABSTRACT:
FIELD OF THE INVENTION
The present invention relates to networks and, more particularly, to transport of data between interconnected networks.
BACKGROUND OF THE INVENTION
As network communications grow more common, it becomes increasingly important to use network bandwidth in the most efficient way possible. In general, data communication through multiple networks is accomplished in one of two ways: routing each portion of the data through the network, called a routing system, or establishing a complete path through the network, called a connection. Each of these systems has its own advantages and disadvantages.
Routing System Transport
A routing system typically consists of tables of data that describe which computers are directly accessible to all other computers. The routing system examines a header within the packet to determine the destination address of a packet and which node in the network to transmit the packet to next, i.e. the destination of the next hop in the path from the source to the destination nodes. Ultimately, the data packet reaches the destination by being examined by each node in the path to the destination. The header within the packet contains an identifier that can be used to determine some of the characteristics of the packet. The resulting overhead in determining the paths on the routing system can reduce performance of data transport from the source to the destination. Consequently, a routing system works well if the data to be transported is relatively small or is concentrated at one point in time. One example of a routing system is transport control protocol/internet protocol (“TCP/IP” ). TCP/IP does not utilize any specific predetermined path from a message source to a message destination. In TCP/IP, messages are routed based upon a destination address contained in the message. The message may take any path to the destination which is available. The message may also contain the address of the source of the message but typically does not contain the address of the devices in the path from the source to the destination. These addresses are determined by each device in the path based upon the destination address and the paths available.
FIG. 1
illustrates a typical interconnection of networks. As seen in
FIG. 1
, a source node
10
is connected to a destination node
35
by a number of networks,
12
,
13
,
14
,
21
,
22
, and
31
and routing devices
15
,
20
,
25
, and
30
. When operating using TCP/IP, source node
10
sends a TCP/IP datagram to the destination node
35
by forwarding the datagram through a series of routing devices. Each subsequent routing device has information regarding networks and other routing devices that may be used to forward datagrams. The routing device
15
forwards the datagram based upon a destination address contained in the datagram by looking up in a routing table the address of the device where the datagram is to be sent. This routing table associates destination addresses with the addresses of accessible devices. Thus, in the present example, if the datagram from node
10
has a destination address specifying node
35
, then routing device
15
may direct the datagram across network
13
to routing device
25
, which in turn would route the datagram over network
22
to node
30
, which would finally route the datagram to the destination node
35
over network
31
. Each step in the path of the datagram from node
10
is referred to as a “hop”. A routing table specifies the address of the next hop in a datagram's path based upon the destination address of the datagram. Thus, the routing table includes the address of accessible devices to the routing device and the destination addresses which are associated with those accessible devices.
The manner in which a routing table is utilized by a routing device is known as a “routing algorithm”. In TCP/IP, the routing algorithm is typically a three step process where the first step is to determine from the destination IP address of the datagram whether the address appears among the direct routes (host devices or networks attach directly to the routing device) specified in the routing table. If such is the case, then the datagram is sent to the directly attached network. If the destination address does not specify a direct route, then the routing device determines if an indirect route (a route through a gateway to a device or network not directly connected to the routing device) is specified for the destination address. If such is the case, then the datagram is forwarded to the specified gateway IP address. Finally, if no direct or indirect route is specified, then the datagram may be sent to a default address or, if no default address is specified, an error message is returned to the source. The routing tables used by routing devices in TCP/IP networks may be arrived at in a number of ways. In one type of routing, a routing table is broadcast to each routing device from the device which maintains a routing table of all possible routes. In other forms of routing, each routing device builds a routing table via its own dialog which is called the routing protocol. These tables are then built and maintained by each routing device. In a third form of routing, a static routing table is used for each message routed.
Each of these cases may have its own disadvantages. For example, the broadcast and routing protocols are repeated on a routine basis to assure the routing device has an accurate routing table. Consequently, this broadcast may occupy network bandwidth. The static routing table, while not requiring network bandwidth, may be inefficient in determining the optimal routing path or, in a worse case, be unable to adapt to changes in the network topography which makes routes unavailable.
With respect to the dynamic routing methods, as the number of interconnected networks and number of devices connected to those networks increases, the size of the routing table and therefore the amount of network resources occupied to maintain the routing table may also increase. In certain circumstances, this overhead may occupy so much of the resources of the network as to cause unacceptable reductions in network performance. Thus, in a routing system, communications are individually routed which incurs overhead of routing each datagram at each network node. Routing may, therefore, be less than optimal for long communications between the same sources and destination nodes. There are, however, certain advantages to a routing system. Specifically, a routing system may have the advantages of reliability and bandwidth utilization. For example, a routing system may achieve increased reliability by its ability to route a message from a source node to a destination node along different or alternative paths. Furthermore, a routing system may achieve greater bandwidth utilization by using only that bandwidth which is necessary to transport the messages required.
Connection Oriented Transport
A connection oriented system consists of virtual connections that link source nodes to destination nodes. The virtual connections are configured early and exist until the data transmission is complete. Data is transported to a destination node by examining the header that indicates which virtual connection to use. Unlike a routing system, nodes in a connection oriented transport do not need to examine the packet to determine which node to route the data to next. Rather the complete connection is established in advance and maintained until the data transmission is complete. The connection oriented transport, however, allocates a discrete portion of the network's bandwidth for each virtual connection. If a virtual connection is not used properly, bandwidth may be wasted. A connection oriented transport works well if a large amount of data is to be transported.
One example of a connection oriented transport is asynchronous transport mode (ATM). ATM transports data by establishing virtual connections between the source and destination nodes. Unlike the routing systems described above, data trans
International Business Machines - Corporation
Myers Bigel Sibley & Sajovec P.A.
Phan M.
Vu Huy D.
Woods Gerald R.
LandOfFree
Traffic flow cutover to virtual connection transport does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Traffic flow cutover to virtual connection transport, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Traffic flow cutover to virtual connection transport will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2529311