Information
-
Patent Application
-
20020001313
-
Publication Number
20020001313
-
Date Filed
April 19, 200123 years ago
-
Date Published
January 03, 200222 years ago
-
Inventors
-
Original Assignees
-
CPC
-
US Classifications
-
International Classifications
Abstract
Data transmission system for transmitting packets of data from a source workstation (10) to a destination workstation (18) wherein the packets of data are transmitted over at least an IP network (14) between an ingress node (16) connected to the source workstation and an egress node (20) connected to the destination workstation, wherein each router within the intermediary nodes along the data path from the ingress node to the egress node determines the best route in a routing table defined by the contents of a field contained in each packet of data being received. For this, the router of the ingress node comprises a configuration table which defines the contents of the TOS field in function of information associated with the 4/5 level protocol such as TCP or UDP.
Description
TECHNICAL FIELD
[0001] The invention relates to IP data transmission networks wherein the route is determined in each router of the data path by using a combination of metrics, and relates in particular to an IP data transmission network that selects routes based on level 4/5 information.
BACKGROUND
[0002] When data packets are transmitted from a source workstation to a destination workstation over an IP data transmission network, the packets are routed from node to node in the network by a routing mechanism implemented by a router in each node of the data path.
[0003] Each IP datagram received by a node and which specifies a destination address other than the local node address is subject to the IP routing algorithm by the router of the node which selects the next node for the datagram. For this, the router uses a routing table which contains information about the other routers of its own network and about IP networks to which its own network is attached.
[0004] The routing mechanism enables the optimal routing path to be determined. Such a path determination is based on a variety of metrics or a combination of metrics, such metrics being values resulting from algorithmic computations on a particular variable or values directly input by the router administrator. The comparison of the metrics allows the router to determine the optimal routes and therefore to establish the routing table.
[0005] Many different metrics have been used in routing algorithms. Sophisticated routing algorithms can base the route selection on multiple metrics, combining them in a manner resulting in a single metric. For this, several metrics may be used. Path length is the most common metric. Some routing protocols allow the network administrator to assign arbitrary costs to each network link. In this case, the path length is the sum of the costs associated with each link traversed. Another important metric to be used is the communication cost, insofar as some companies may not care about performance whereas they care about operating expenditures. Other metrics may also be used, such as the reliability (usually described in terms of bit error rate) of each network link, delay (the time required to move a packet from a source to a destination), bandwidth (available traffic capacity of a link), or load (the degree to which a network resource such a router is busy).
[0006] All the information used by the routing protocol to determine the metrics comes from the lowest three layers of the protocol stack. It is important to note that in the internet protocols as well as in the Open System Interconnect (OSI) model, the layer which defines the packet delivery, including routing, is the third layer. Put conversely, the upper layers of the protocol stack, which constitute the layer 4/5 in the network procedures, are not used to determine metrics and therefore to build the routing table. These upper layers define the application such as the Transmission Control Protocol (TCP) or the User Datagram Protocol (UDP).
[0007] There have been several attemps to use the level 4/5 information in the routing algorithm. All these approaches were based on the assumption that the source and destination workstations provide information to the routers and support a specific protocol to set the inputs. These attempts, which focused on using level 4/5 information (for the TCP/IP model as well as for the OSI model),have been abandoned, as the did not result in stable and durable implementations. One such attempt was implemented in OSPF, using the configuration parameters called Type of Service (TOS). This mechanism requires the application to set the TOS field in all IP data generated by the source workstation. Unfortunately, the support of TOS in routing requires modifications to the application. Today, no application is implemented in such a way. Therefore, the use of the TOS field has been dropped in recent OSPF implementations according to the latest RFC recommendations. Some other attempts use a priority queue in the routers: the router tries to look at some field significant of level 4/5 application layer when receiving the IP data packets, in order to set some priority in an input/output buffer. Such a mechanism acts on the data transmission speed, but does not modify the routing path.
SUMMARY OF THE INVENTION
[0008] Accordingly, the main object of the present invention is to provide a route selection in the routing algorithm of an IP data transmission network which depends upon the type of application used in a communication between a source workstation and a destination workstation.
[0009] Another object of the present invention is to select a routing path in OSPF routing protocol according to the TCP or UDP port required in each router of a transmission path in an IP data transmission network.
[0010] The invention relates therefore to a data transmission system for transmitting packets of data from a source workstation to a destination workstation. Packets of data are transmitted over a communication network such as an IP network between an ingress node connected to the source workstation and an egress node connected to the destination workstation. Each router within the intermediary nodes along the data path from the ingress node to the egress node determines the best route in a routing table defined by the contents of a field contained in each packet of data being received. To accomplish this, the router at the ingress node comprises a configuration table which associates the contents of the field with information carried by a protocol layer above the IP level, generally the 4/5 level such as TCP or UDP. In one embodiment of the invention, the field employed is the Type-of-Service (TOS) field.
BRIEF DESCRIPTION OF THE DRAWINGS
[0011] The above and other objects, features and advantages of the invention will be better understood by reading the following description of the invention in conjunction with the accompanying drawings wherein:
[0012]
FIG. 1 is a block-diagram representing a data transmission network wherein the invention may be implemented.
[0013]
FIG. 2 is a schematic representation of an IP datagram with the various fields of the IP header.
[0014]
FIG. 3 shows the contents of the service type field of the IP datagram including the TOS bits.
[0015]
FIG. 4 is a schematic representation of the IP datagram showing the TCP/UDP header after the IP header.
[0016]
FIG. 5 is a flow-chart of the operation of an ingress node according to the invention.
[0017]
FIG. 6 is a schematic representation of configuration table used in the data transmission system according to the invention.
[0018]
FIG. 7 is a block-diagram representing the router of the ingress node according to the invention.
DETAILED DESCRIPTION OF THE INVENTION
[0019] A data communication system that is suitable for accomodating the present invention is shown in FIG. 1. As a source workstation 10 attached to a LAN 12 may access an IP network 14 through an ingress node 16, in order to transmit packets of data to a destination workstation 18 that is connected to the network 14 by an egress node 20.
[0020] The optimal route, which is found by computing a combination of metrics, is determined in the ingress node 20 according to the level 4/5 information as described below. With one application, such a route could be through a first intermediary node 22, whereas with a different application the optimal route could be through a second intermediary node 24.
[0021] For example, suppose that the first application is a Voice-over-Internet-Protocol session (VoIP) requiring low delay, but not requiring a large bandwidth, and suppose that the second application is a data batch transfer using File Transfer Protocol (FTP), without any particular delay requirement but requiring large throughput. Further, suppose that the delay through the first intermediate node 22 is less than the delay though the second intermediate node 24 but, the overall bandwidth on the former route is limited. Also, suppose that the delay through the second intermediate node 24 is greater but a lot of free bandwidth is available on this second route. Because of the shorter delay on the route through the first intermediate node 22, both sessions, voice and data, would be established on this route when using legacy OSPF, whereas the invention based on the 4/5 level information enables two optimal routes to be established, thereby improving the load balancing between all routes in the network.
[0022] As illustrated in FIG. 2, an IP datagram data includes an IP header 24 containing the information necessary to send the packet in correct form over the route, such as the source IP address or the IP destination address, and the IP data field 26. The header 24 includes a service type field 28, which is illustrated in FIG. 3. Such a service type field includes two bits for precedence (a measure of the priority of the datagram), a MBZ bit (must be zero) reserved for future use, and four Type Of Service bits (bits 3-6), which characterize some parameters of the application such as the propagation time, the throughput, and the reliability.
[0023] As shown in FIG. 4, the IP data field 26 of an IP datagram includes a header TCP/UDP and data. Note that in both TCP and UDP protocols and other 4/5 level protocols, the header includes a field containing the source port number and a field containing the destination port number.
[0024]
FIG. 5 shows the operation of the ingress node 16 (see FIG. 1), when a frame is received (step 30). First, the protocol identification is made by extracting (step 32) the associated field in the IP header 24 of the IP datagram (see FIG. 2). The source and/or destination port number is also extracted (step 34) from the 4/5 level header. Then, there is a lookup of a configuration table represented in FIG. 6 (step 36). Such a configuration table provides the TOS field corresponding to a given protocol and a given port number (source and/or destination). It must be noted that, insofar as there can be numerous port numbers for each protocol being identified, such a table may not contain all possible cases.
[0025] The lookup of the configuration table determines whether an entry of the table is associated with the protocol being identified (step 38). If so, it is then determined whether the port number extracted from the 4/5 level header is identified in the configuration table (step 40). If the port number is found, the TOS bits in the IP datagram are replaced (step 42) by the bits which are found in the configuration table. It must be noted that the TOS bits are previously all zeros corresponding to the default route. Finally, the routing task is achieved by the OSPF protocol or an equivalent protocol (step 44) before transmitting the frame over the network.
[0026] When the protocol being identified in the IP datagram does not correspond to any entry in the configuration table, or when the port number for this protocol is not found in the configuration table corresponding to the port number identified in the 4/5 level header, the TOS bits are set to zero or not changed if they were already set to zero (step 46). Then, the OSPF (or equivalent protocol) routing task is achieved by the router (step 44). Note that, in the latter case, the “all zeros” TOS defines only the default route.
[0027] As illustrated in FIG. 7, the router of ingress node 20 (see FIG. 1) includes a protocol processing unit 50 which identifies the protocol associated with the application in the received frame. Thus, such a protocol can be UDP plus Real Time Protocol (RTP) for the voice flow, or TCP plus FTP for the data flow. Note that the protocol processing unit 50 also looks for the port number associated with the protocol in the frame header as already mentioned.
[0028] When the protocol and the port number are identified by the protocol processing unit 50, the frame-together with this information is transmitted to a forwarding processing unit 52. The protocol and the port number enable the forwarding processing unit 52 to look in the configuration table 54 for determining the value of TOS bits corresponding to this protocol and this port number. Then, the forwarding processing unit 52 replaces in the frame the previous TOS bits (generally zero bits) by the value determined in the configuration table 54, and stores the frame into a frame buffer 56. The TOS value also enables the forwarding processing unit 52 to select an appropriate routing table identifying the route to be used, for example a routing table 58 containing routing information for the UDP protocol, or a routing table 60 containing routing information for the TCP protocol. Then, the forwarding processing unit 52 utilizes this routing information to select an output queue for transmitting the frame over the network. In the previous example, the voice frame using UDP will have routing information that will enable the transfer of the voice frame from the frame buffer 56 to an output queue 62 for transmission to the first intermediate node 22, whereas the data frame using TCP will have routing information that will enable the transfer of the data frame from the frame buffer 56 to an output queue 64 for transmission to the second intermediate node 24.
[0029] The routers of all the other nodes of the route such as the first and second intermediate nodes 22 and 24 include the same components except the configuration table, since these nodes have no need to determine and to change the value of the TOS field in the frame. In the same way as in the ingress node, the TOS value enables each router to select the appropriate routing table in order to know the routing information and to transmit the frame to the next node of the route.
Claims
- 1. A router suitable for use in transmitting a packet of data through a communication network wherein the best route through the network is determined at each node, said router comprising:
logic for identifying and extracting higher-layer information carried by at least one of the layers above the lowest three layers of a communication protocol of a received Packet; a configuration table for associating the higher-layer information with lower-layer information carried by at least one of the lowest three layers of the communication protocol; and a routing table for determining routing of the packet, responsive to the lower-layer information.
- 2. The router of claim 1, wherein the higher-layer information comprises a protocol identifier and a port number.
- 3. The router of claim 1, wherein the lower-layer information comprises a type of service identifier.
- 4. The router of claim 3, wherein the higher-layer information comprises a protocol identifier and a port number.
- 5. The router of claim 1, wherein the lower-layer information is carried by Internet Protocol (IP).
- 6. The router of claim 1, wherein the higher-layer information is carried by Transmission Control Protocol (TCP).
- 7. The router of claim 1, wherein the logic comprises a protocol processing unit.
- 8. The router of claim 7, wherein the logic further comprises a forwarding processing unit.
- 9. A method for determining the route of a packet through a communication network, said method comprising the acts of:
a) extracting higher-layer information carried by at least one of the layers above the lowest three layers of a communication protocol of a packet; b) associating the higher-layer information with lower-layer information carried by at least one of the lowest three layers of the communication protocol, c) using the lower-layer information to select a route for the packet through the network by accessing a routing table containing a plurality of routes.
- 10. A method for determining the route of a packet through a communication network, said method comprising the acts of:
a) receiving a packet; b) identifying and extracting higher-layer information carried by at least one of the layers above the lowest three layers of a communication protocol of the received packet; c) associating the higher-layer information with lower-layer information carried by at least one of the lowest three layers of the communication protocol by accessing a configuration table; And d) determining routing of the packet by accessing a routing table responsive to the lower-layer information.
- 11. The method of claim 10, wherein the higher-layer information comprises a protocol identifier and a port number.
- 12. The method of claim 10, wherein the lower-layer information comprises a type of service identifier.
- 13. The method of claim 12, wherein the higher-layer information comprises a protocol identifier and a port number.
- 14. The method of claim 10, wherein the lower-layer information is carried by Internet Protocol (IP).
- 15. The method of claim 10, wherein the higher-layer information is carried by Transmission Control Protocol (TCP).
Priority Claims (1)
Number |
Date |
Country |
Kind |
00480053.8 |
Jun 2000 |
EP |
|