Multiplex communications – Network configuration determination – Using a particular learning algorithm or technique
Reexamination Certificate
2000-10-05
2004-10-12
Ton, Dang (Department: 2661)
Multiplex communications
Network configuration determination
Using a particular learning algorithm or technique
C370S278000, C370S395210, C370S396000, C370S400000, C709S226000
Reexamination Certificate
active
06804201
ABSTRACT:
BACKGROUND OF THE INVENTION
This invention relates to a cognitive packet network and method of packet switching packets in a data packet communication environment having a plurality of digital packet transmission stations interconnectable in paths.
Most communication networks have been arranged to provide either a connection-oriented service or a connectionless service. In a connection-oriented service, a circuit is established between a source node and a destination node of the network, and all the data transferred between the source node and destination node travels via the links and nodes incorporated in the circuit. The circuit is released only when the call is terminated. Such an arrangement has the advantage that all the data arrives at the destination node in the same order as it is transmitted from the source node. In a connectionless service, data is transmitted between a source node and a destination node via packets or datagrams which each include the address of the destination node and travel along one or more paths through the network to the destination node, the paths chosen generally depend upon the conditions in the network. Accordingly, the packets from the source node can reach the destination node via a multiplicity of paths through the network, each having different delays so that the order in which the packets are received at the destination node does not necessarily correspond to the order in which the packets are transmitted from the source node. Accordingly, facilities are normally required to reorder the data receive by a destination node of such a network.
A wide area network (WAN) is sometimes referred to as a “network of networks”. The Internet is a WAN that has become popular.
ARPANET (Advanced Research Projects Agency Network) was the first network in the USA to use packet switching technology. It was set up as an experimental network to investigate system resource sharing in 1969, and has given rise to the Internet which currently connects nodes throughout the world. Routes are generated in the Internet in a distributed fashion. Each node maintains a network database, which includes a complete network topology. The Internet uses datagrams, so that circuits do not have to be construed. Each node calculates a Routing Table (RT) from its delay table and topology database; this routing table contains the best outgoing link to all destination nodes in the network. An end-to-end protocol known as TCP is used to ensure that packets are received at the destination in the correct order, and that losses are detected.
The Internet is based upon a transmission protocol known as “Transmission Control Protocol/Internet Protocol” (or “TCP/IP” for short), which sends packets of data between a host machine, e.g. a sever computer on the Internet, and a client machine, e.g. a user's personal computer connected to the Internet. The World-Wide-Web (WWW) is an Internet interface protocol which is supported by the same TCP/IP transmission protocol.
The following U.S. patents relate to network communications: Prior U.S. Pat. No. 6,038,230 relates to time management in a system in which packets are transmitted only at predetermined clocked instants of time (“pre-defined periodic intervals”), as when dealing with the transmission of voice packets over a network. It describes a scheme for automatically inserting packets into the first available subsequent interval on each pre-determined link, whenever packets are delayed and thus cannot be transmitted within their predetermined time interval.
In contrast, the present invention does not assume that there are any pre-defined periodic time intervals for packet transmission but deals with packet routing through the network rather than with the selection of specific time intervals for packet transmission. It determines adaptively the sequence of nodes and links that a packet must traverse.
In U.S. Pat. No. 6,044,062 to Brownrigg et al., a routing scheme for a wireless network can be used to connect to the Internet. The purpose of the system is to allow “clients” to communicate via wireless with a “server” using data packets, either directly (one hop), or indirectly by communicating via other “clients”. It optimizes the communication by seeking to minimize the number of hops, or it can also factor in traffic load and link reliability issues. Each client creates and maintains a route (path) to the server. The routing algorithm portion of this patent specifies that if a client cannot reach a server directly, it probes all of its neighbors. As it does so, if any neighbor has an existing route to a server, then this information will propagate to the server which will add the new client to its routing table. Information about the route can then be forwarded from the immediate neighbor down to the new client so that both the server and the client have the route. The routing scheme proposed here is very similar to the one in U.S. Pat. No. 5,142,531, below, as adapted to the particular wireless environment. Thus the present patent's contribution is related specifically to the broadcasting wireless environment, rather than to routing issues.
U.S. Pat. No. 5,142,531 to Kirby proposes that each node N
S
in a packet switching network will contain a table where each row in the table corresponds to a possible destination node, say N
A
. Each row in the table has two entries. Consider the row corresponding to destination node N
A
:
the first entry is the name of the immediate neighbor node of N
S
which has an ultimate connection to destination N
A
with the minimum number of hops (intermediate nodes),
the second entry is the name of the immediate neighbor node of which provides the currently known shortest delay from N
S
to N
A
.
The method uses both lists to build a route from source to destination in advance of the actual transmission of a sequence of packets from source to destination. In order to do this, the source sends out a BUILD packet which is routed hop-by-hop (node-by-node) from source to destination according to a very specific scheme.
At any step of BUILD's route, if the next node designated by List 1 and List 2 is the same, then it is selected as the next node. If the two are not identical, then BUILD goes indirectly (i.e. in more than one step) to the node indicated in List 1 which is now included in the route, but first uses the link (next node) designated in the appropriate List 2.
In addition to this specific algorithm used by the BUILD packet, this patent also proposes a manner in which the contents of List 1 and List 2 are to be maintained and updated. All nodes are organized so as to propagate hop count and delay information to all other nodes in a periodic fashion.
Each neighboring node will pass on to its neighbors its known hop count information to its possible destinations, while the receiving neighbor will simply add 1 to that hop count to deduce its own hop count distance to the possible destinations. Presumably the identity of the neighbor having minimum hop count to destination, which is the information stored in List 1, will be selected from all the available values. Delay information will similarly be aggregated by propagating delay information from neighbors and adding to it the link delay between adjacent neighbors. Similarly, the neighbor providing the minimum delay to a destination should presumably be entered into the appropriate row in List 2.
The patent descriptions also indicates that the information communication between nodes will only take place when a nodes own List 1 or List 2 change, and in addition to that a node will confirm periodically to all of its neighbors that it “exists”—i.e. that it is still operational.
In addition to this, when a route is established by a BUILD packet, and the BUILD packet reaches the ultimate destination, and ACKNOWLEDGE packet is sent back to the source so that the source can know that the route is established so that the normal transmission of packets can begin.
In contrast, the present invention uses “cognitive” or “smart” packets to search for ro
Hobby, III William M.
Mehra Inder Pal
Ton Dang
LandOfFree
Cognitive packet network does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Cognitive packet network, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Cognitive packet network will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3285402