Patent Application
20070242607
The present invention generally relates to distributed source routed networks and more particularly to methods and systems for controlling distribution of topology information between nodes within a distributed source routing network.
Data communications networks include a variety of network elements, such as computers, servers, nodes, routers, switches, bridges, clubs, proxies and other devices coupled to, and configured to pass data between, one another. In general, the data is formatted and passed as data units between the network elements through the use of one or more communication links and in accordance with a predetermined protocol, such as Internet Protocol packets, Ethernet frames, data cells, segments and other data formats. The network elements may have unique addresses (e.g., IP addresses) assigned thereto. A particular protocol for the data unit may be handled by multiple network elements and supported across multiple communications links as the data units are routed over the network between source and destination addresses.
Most protocols fall into one of two categories, namely Distance Vector protocols and Link State protocols. Distance vector protocols in general determine the distance between source and destination nodes by calculating the number of hops a packet traverses en route from the source node to the destination node. Distance vector protocols (e.g., Bellman-Ford algorithms) call for each node to send all or some portion of the node's routing table to the neighboring nodes. For each network path, the receiving node picks the neighboring nodes that exhibits the lowest cost and adds the neighboring nodes to its routing table for re-advertisement. Examples of distance vector protocols are RIP and RIP-2 algorithms. In general, Link State protocols send small updates throughout a network, while distance vector protocols send larger updates only to neighboring nodes. In distance vector protocols, the nodes know only about neighbors nodes. Link state protocols converge more quickly, and thus are somewhat less prone to routing loops than distance vector protocols. Also, while link state protocols have utilized more CPU power and memory as compared to distance vector protocols, link state protocols are generally more scalable than distance vector protocols.
Nodes within the network cooperate and communicate with one another to convey the data units between network elements. For example, an originating network element may provide a destination IP address to a source node. The source node uses the destination IP address, among other things, to calculate a route to another node that is joined to the network element having the destination IP address. The route traverses through one or more nodes within the network. In conventional source routing networks, the source node calculates the route based on routing tables that are stored at the source node. The routing tables define an overall network topology map for the network. Each node stores identical routing tables and thus each node has the same “view” of the overall network topology map. The source node may utilize various different metrics when implementing routing algorithms to determine the best route. Sophisticated routing algorithms can base route selection on multiple metrics.
Source routing, in general, represents techniques whereby a source node of a packet specifies a route that a packet should take through a network area as a packet travels through the network. In source routing, the source node in general makes some or all of the decisions as to which nodes a packet should travel through en route to the destination. In certain source routing implementations, the source node specifies the exact route along which the packet is to travel all the way to the destination IP address. In other source routing implementations, (e.g. lose source record routing (LSSR)), the source node provides one or more of the nodes through which the packet should travel, but not necessarily the entire path. Examples of link state algorithms include Open Shortest Path First (OSPF), intermediate system to intermediate system (IS-IS), and Netware's Link Services Protocol (NLSP).
In source routing algorithms, routing information is flooded to all nodes in the network in order that all nodes shall have an entire view of the overall network topology. As link states change, each node sends updated link state information, as link state advertisements, over the network to all other nodes. The link state advertisements describe the connectivity associated with the sending node. The link state information also includes link state attributes associated with each link state. Each node builds and maintains routing tables for the overall network topology map of the entire network.
However, existing distributed source routing implementations have experienced certain limitations. In conventional link state routing protocols in support of hop-by-hop routing, all nodes within a network must have a common overall network topology map in order to guarantee loop free routing. If each node was not provided with a common overall network topology map, routing loops would occur whereby nodes along a routing path might pass a data unit back to an upstream, preceding node that has already passed the same data unit. Thus, the data unit may fall into a never ending loop. To maintain the overall network topology map at each node, substantial detailed link state information is shared between all nodes. As link state information changes at a given node, the node updates the rest of the network on a dynamic basis through link state advertisements.
In certain environments, it may be undesirable to provide substantial detailed link state information about all nodes to one another. For example, different nodes within a network may be owned, managed or controlled by different entities (e.g., service providers and the like). Therefore, once each node constructs an overall network topology map, the entity that controls the node will have access to detailed link state information about the network capabilities of another entity. As another example, when capacity is being added to a network, it may be desirable to limit access initially the added capacity. For example, when new nodes, more bandwidth, new links (e.g. additional fiber connections) and the like are added to a network, it may be desirable to test fully the added capacity before making such capacity fully available to the network. Heretofore, when capacity was added to a network, the capacity was placed on the network all at one time, namely the capacity was either entirely unavailable or entirely available to all routing control elements in the network.
A need remains for improved methods and systems for controlling distribution of network topology information between nodes of a network that utilize a distributed source routing protocol.
In accordance with one aspect of the present invention, a network topology management system is provided for managing distribution of network topology information between nodes within a network that performs a distributed source routing protocol. The system includes nodes, distributed throughout a network, that are interconnected by node-to-node links, each of which is described by associated link state information. The system includes a control module that is joined to the nodes and that receives the link state information from each node. The control module records the link state information in master routing tables to form an overall topology map of the network. The control module stores predetermined policy statements that define a manner in which the link state information should be distributed between the nodes. The control module modifies the link state information to form modified link state information that yields abridged routing tables based on the predetermined policy statements. The abridged routing tables differ from the master routing tables of the overall network topology map. The control module distributes the modified link state information to at least one of the nodes in the network.
Optionally, the control module may produce multiple different sets of abridged routing tables, were each set of abridged routing tables forms a unique abridged network topology map that differs from other abridged network topology maps and that differs from the overall network topology map. The link state information may include link states and link state attributes, at least one of which is modified by the control module to create the abridged routing tables. Optionally, the control module may entirely mask a select link state such that an abridged routing table entirely excludes connectivity between a select node and a select neighboring node. Optionally, the control module may modify link state attributes to form different first and second sets of abridged routing tables that are provided to different first and second nodes, respectively, such that the first and second nodes have different link state information associated with a common select link state.
In accordance with an alternative embodiment, a method is provided for controlling distribution of topology information in a network that utilizes a source routing. The method includes receiving link state information associated with links between nodes within a network and recording the link state information in master routing tables to form an overall network topology map of links within the network. The method includes modifies the link state information to form modified link state information that yields abridged routing tables based on predetermined policy statements, wherein abridged routing tables differ from the master routing tables of the overall network topology map. The method further includes distributing the modified link state information to at least one node in the network.
The following detailed description sets forth certain examples to provide an understanding of embodiments of the present invention. The functions described herein may be implemented as a set of program instructions that are stored in a computer readable memory within one or more network elements and executed on one or more processors within the network elements. It is also understood that the various functions described herein may be implemented using integrated circuitry, such as application-specific integrated circuits (ASIC), programmable logic used in conjunction with a programmable logic device such as a field programmable gate array (FPGA), microprocessor, state machine or any other device including any combination thereof. Programmable logic may be fixed temporarily or permanently in a tangible medium such as in a read-only memory, random-access memory, a memory stick or other storage medium. Programmable logic may also be fixed in a computer data signal embodied in the carrier wave, allowing the programmable logic to be transmitted over an interface such as a computer bus or communications network.
The network 10 also includes a control topology (e.g. a hub topology) at the control level or plane. The control topology is shown by the dotted lines 18 between each of the nodes 12 and a designated topology management module 16. In
The nodes 12 establish connectivity over the links 14, wherein each connection is defined by a link state and a series of link state attributes (collectively “link state information”). The link state attributes are utilized as various metrics by the source routing algorithm in connection with calculating routes through the network 10. For example, the following link state attributes may be used, namely path length, reliability, routing delay, bandwidth, load, communication cost, geographic location, diversity routing and the like. “Path length” is a common routing metric. Some routing protocols allow network administrators to assign arbitrary costs to each network link, in which case path link represents the sum of the costs associated with each link traversed. Other routing protocols define hop count, a metric that specifies the number of passes through internetworking nodes that a packet takes en route from a source to a destination. “Reliability” refers to the dependability (usually described in terms of the bit-error rate) of each network link. Some network links might go down more often than others. When a network fails, certain network links might be repaired more easily or more quickly than other links. Reliability factors may be taken into account in the assignment of the reliability ratings, which are arbitrary numeric values usually assigned to network links by network administrators. “Routing delay” refers to the length of time required to move a packet from source to destination through the network. Delay depends on many factors, including the bandwidth of intermediate network links, the port queues at each router along the way, network congestion on all intermediate network links, and the physical distance to be traveled. Because delay is a conglomeration of several important variables, it is a common and useful metric.
“Bandwidth” refers to the available traffic capacity of a link. For example, all other things being equal, a 10-Mbps Ethernet link would be preferable to a 64-kbps leased line. Although bandwidth is a rating of the maximum attainable throughput on a link, routes through links with greater bandwidth do not necessarily provide better routes than routes through slower links. For example, if a faster link is busier, the actual time required to send a packet to the destination could be greater. “Load” refers to the degree to which a network resource, such as a router, is busy. Load can be calculated in a variety of ways, including CPU utilization and packets processed per second. Monitoring these parameters on a continual basis can be resource-intensive itself. “Communication cost” may be used, such as when certain companies may not care about performance as much as they care about operating expenditures. Although line delay may be longer, a company may choose to send packets over their own lines rather than through the public lines that cost money for usage time.
Once the nodes 12 establish the connections over the links 14, the nodes 12 supply link state information to the topology management module 16 over the control link 18. The link state information includes, among other things, the link state and the link state attributes associated with each link 14. The topology management module 16 receives link state information, such as link states and link state attributes) that is associated with each link 14 between nodes 12 within the network 10. The link states and link state attributes collectively, define an overall topology view of the network. The topology management module 16 combines the link state information to form a set of master routing tables that define an overall network topology map. The topology management module 16 also stores a table of predetermined policy statements. The policy statements include a series of rules that are associated with select nodes, link states and link state attributes. The rules filter or regulate distribution of the link state information to the nodes 12. The module 16 uses the policy statements to modify the link state information to form modified link state information that yields or create sets of abridged routing tables, where each set of abridged routing tables differs from one another and from the master routing tables. The abridged routing tables may be constructed at the module 16 before distribution to the nodes 12. Alternatively, the module 16 may distribute the modified link state information to the nodes 12 that then construct the abridged routing tables. In both of the above examples, the modified link state information “yields” the abridged routing tables regardless of whether the abridged routing tables are constructed at the module 16 or at the node 12. The topology management module 16 distributes the abridged routing tables or modified link state information to the nodes 12 in the network 10 based on the policy statements.
When the module 16 creates multiple different sets of abridged routing tables, each abridged routing table provides different “view” of link state information to a corresponding node. The module 16 may create a set of abridged routing tables by modifying link state information associated with a select node 12. The module 16 may also create an abridged routing table by entirely masking one or more link states such that the abridged routing table excludes entirely one or more node-to-node links associated with the masked link state.
The nodes 12 convey data units between one another based on source routing calculations. Each node 12 performs source routing calculations based on the abridged routing tables that are stored locally at the node 12.
The control module 20, records the link state information from each node 12 in the master routing tables to construct an overall network topology map 24. The overall network topology map 24 may be formed as a series of routing tables 50, each routing table 50 of which is associated with a corresponding node 12. Each routing table 50 may include a link state field 52 for link states, over which connections have been established, for a node 12.
In the example of
The module 16 stores, in memory 26, a table of policy statements 28. The policy statements 28 are used by the control module 20 in connection with controlling distribution of link state information from the routing tables 50 to nodes 12. The policy statements 28 may be provided in a variety of forms. In the example of
In the example of
The rules 156b, 156g and 156h provide examples of how the policy statements 128 may be utilized to modify link state information from the original form and provide the modified link state information to the nodes 12. Rule 156b indicates that the link state attribute associated with bandwidth for the link state #6-#9 is to be modified or abridge from an actual value or condition to reduce by 50% the indicated bandwidth capacity of the corresponding link state. Thus, rule 156b directs the control module 20 to create a “pseudo value” for the bandwidth attribute of the link state #6-#9. For example, if the full bandwidth capacity of link state #6-#9 represents 10 Mbps, the control module 20 will abridge the actual bandwidth information and instead inform each node 12 within the network 10 that the pseudo bandwidth capacity associated with links state #6-#9 is only 5 Mbps.
Rule 156g instructs the control module 20 to modify the link state attribute, associated with the weight or cost of link state #4-#2, from an actual value by doubling the weight associated with the link state to a pseudo value. For example, if the actual weight associated with link state #4-#2 represented 100, the control module 20 abridges the weight information and informs each node 12 within the network 10 that the pseudo weight associated with link state #4-#2 was 200.
Rule 156d provides an example for when it may be desirable to entirely hide or mask connectivity between two nodes. Rule 156d corresponds to the link state #6-#8 between nodes #6 and #8, and instructs the control module 20 to entirely hide or mask from all other nodes 12 the existence of the link state #6-#8. Returning to
Optionally the rules within rule field 164 may be correlated, not just too specific link state attributes within link state attribute field 162. In addition, the specific rules may be directly correlated to target nodes that are to receive an abridged routing table. For example, rule 156h instructs the control module 20 to reduce by 50% the bandwidth attribute associated with link state #4-#9 before providing the bandwidth attribute to target node #5. In this example, all other nodes #1-#3 and #6-#9 are informed of the full bandwidth capability of the link state #4-#9. Only node #5 is provided with pseudo bandwidth information, namely a 50% reduction from the actual bandwidth value. Rule 156h is specific to a desired target node #5. Thus, the target node #5 will have a different “view” or different link state information than other nodes (e.g., nodes #1-#4 and #6-#9).
The LSI collection module 621 communicates with the nodes 12 (
The topology map reconstruction module 623 accesses the policy statements 28 in memory 26, as well as the routing tables 50 in memory 22, to form the modified link state information that is utilized to construct the abridged topology maps 34 and 44 (
In accordance with certain of the above embodiments, methods and systems are provided in which links states and/or link state attributes may be hidden and are modified, such as to limit traffic over a particular blank. In accordance with certain of the above embodiments, a hub topology is provided to allow control over the distribution of links state information, thereby providing the ability to prevent a link from being reported, two modify links state information and the like. In accordance with certain embodiments, carriers are afforded the ability to control traffic over all or a portion of the network, to prevent certain links from being used, to limit the sharing of information about a particular carriers network topology and the like.
Optionally, the policy statements 28 may be recorded in a form other than a table. Also, the rules may be organized in a manner other than in connection with particular link states and link state attributes. For example, a rule may be associated with a group of link states and/or a group of link state attributes. Further, the rules may be associated with destination nodes, as opposed to a source node. For example, a rule may provide that a particular node (to receive a set of abridged routing tables) is not to be provided certain link state attribute information, independent of the source node with which the link state attributes originated. As a further example, the rules may be organized into classes or levels of priority. For example, the nodes 12 may be classified in different levels, where each level is entitled to abridged routing tables having a different level of detail regarding the link state information of other nodes.
While the above examples are provided in connection with meshed topologies, it is understood that other topologies may be utilized. It is understood that the embodiments described above may be implanted using various routing protocols, such as the OSPF protocol, the intermediate system to intermediate system intra-domain routing exchange protocol and the like.
While the above embodiments are described in connection with link state protocols used during source routing, the present invention may be utilized with non-link state protocols utilized during source routing. For example, the present invention may be used with path vector protocols (e.g., distance vector protocols that include path information) utilized during source routing. The path information includes link state information for a series of links interconnecting nodes used to reach a destination. Similar to the embodiments discussed above, the path information would be used to construct master overall distribution tables.
For example, similar to the process of
While the invention has been described in terms of various specific embodiments, those skilled in the art will recognize that the invention can be practiced with modification within the spirit and scope of the claims.
