A method for transmission of data packets through a network

ES 2 306 830 T3

DESCRIPTION

Procedure for transmitting data packets through a network.

Networks consisting of central nodes and border nodes are considered below. At least some of the central nodes are connected to one or more border nodes, so that each border node is connected to only one central node. The central nodes of the considered network do not have storage capacities. This will be explained in detail later.

The border nodes receive data packets from sources and send data packets to receivers outside the network.

For transmitting data packets through a network, different procedures are known.

At a network border node, such as an Optical Burst Switched (OBS) network, incoming or received data packets, such as Internet Protocol (IP) packets, transfer mode cells Asynchronous Transfer Mode (ATM) or Protocol Data Units (PDUs) are added at the node to a burst with an associated header. The burst contains a number of data packets for a certain destination and is aggregated according to a predefined aggregation strategy. This burst is sent over the network to the destination. Before the burst is sent, a path, path or trail has to be established or reserved in the network so that the burst is properly transmitted through the network to its destination. This is done through the associated header. After the aggregation of the burst, the header is generated and sent from the border node to the connected or associated central node. There the header is evaluated and a path is reserved for the next burst. The header is then sent to a next hub, the header is evaluated, a path is reserved, and the header is sent to a next hub until a path to a destination is reserved for the next burst. After sending the header and after a certain time has elapsed at the border node, which is for path reservation in the central nodes, the burst is sent to the associated central node. The burst is sent from hub to hub through the reserved path to a destination of the burst, which is a certain border node. There the burst is separated into data packets, which are sent to additional destinations.

In networks such as Optical Packet Switching (OPS), the same principles apply as in OBS networks, with the difference that each data packet is transformed into a packet with an associated header. Packages are sent as bursts. The same reservation and transmission principles apply as for the header of the burst for the header of a packet.

In principle, in case of an OBS and OPS network, storage capacities for bursts or packets are not provided in the central nodes of the network. They could be provided by fiber delay lines to store bursts or optical packets. These optical storages are currently very expensive. Limited storage capacities are provided for the header of a burst or packet, since they are much smaller. The headend can be converted from optical to electrical, electrically evaluated, stored and modified, and converted to optical again.

A key feature of these networks is the one-pass reservation scheme of network resources respectively network nodes for each individual burst or packet. Bursts are sent without a successful path set confirmation and burst loss may occur in the event of a dispute.

Examples for OBS networks are described in the following documents. In “Application of Tell & Go and Tell & Wait Reservation Strategies in an Optical Burst Switching Network: a Performance Comparision” by A. Detti, M. Listanti, University of Rome, a comparison between two resource reservation techniques when used in an all-optical network to support high-speed IP traffic.

In "Choices, Features and Issues in Optical Burst Switching" by C. Qiao, M. Yoo, University of Buffalo (SUNY), design choices in burst switching are explored and a new variation is described that is especially suitable for WDM networks. optics. The source nodes of an OBS network are adapted to operate in the electrical and optical domain. According to the Enough Time Only protocol, a source sends a control packet, which is forwarded by a burst after a lag time. The burst is buffered at the source in the electrical domain while the control packet is being processed. If the requested bandwidth is not available, the burst is said to be blocked, and will be sent or can be buffered and then transmitted later.

An architecture for a network is presented in "JumpStart: A Just-in-Time Signaling Architecture for WDMBurst-SwitchedNetworks" by Ilia Baldine, George N. Rouskas, Harry G. Perros, and Dan Stevenson, MCNC and North Carolina State University. Core DWDM using the concept of optical burst switching coupled with a just-in-time signaling scheme.

In a ring network, a number of central nodes are connected to a ring, so that at least part of the central nodes have a connection to at least one associated border node, where a border node is connected to a central node .

ES 2 306 830 T3

Any node that detects a failure is adapted to send a failure message back to the source node (page 89, left column, last paragraph). Bursts belonging to the same connection are guaranteed to take the same path, but they are not guaranteed to be successful.

In a star network, there is only one central node, which has connections to several border nodes.

In a bus network, a number of central nodes are connected in a row, where each central node has two neighbors, except the central nodes at the end of the bus (respectively row). At least a part of the central nodes are connected to at least one border node, so that each border node has a connection only to one central node.

It is an object of the invention to present methods for the transmission of data bursts or packets through a ring, star or bus network without loss of bursts or packets respectively.

This object is achieved by the characteristics cited in claim 1 or 2.

Due to the fact that bursts respectively packets are not sent in case of a lock conflict or reservation respectively dispute in a central node, the advantage of the invention is that losses of bursts or packets are avoided.

Further improvements of the invention are identified in the dependent claims.

In one embodiment of the invention, a response is sent from the central node to the border node with a time or time delay for a central node occupation. Said time or time delay is generated by the central node according to an analysis of its reservation scheme respectively reservation planning for the data bursts or packets. After the analysis, the central node calculates a time or time delay for the association border node, when a new attempt to send a header respectively for the reservation of a path appears to be successful. This time or time delay is inserted into the response and transmitted to the border node. The response is received and evaluated. After this time or time delay elapses, the border node sends the previously sent header back to the central node, to try to reserve a path for its burst or packet. Meanwhile the burst or packet is stored at the border node. This has the advantage of an inexpensive attempt at a reservation.

In another embodiment of the invention, in case of a reservation conflict, the central node schedules a reservation for the burst or packet of the associated header, determines a particular time or time delay for the transmission of said burst or packet, sends a response in the one that is contained said time or particular time delay to the originating border node. There the response is evaluated, the sending of the burst or packet is abandoned, the burst or packet is stored in the border node and after said time or time delay has elapsed, the burst or packet is sent to the central node, to use the planned reservation. This has the advantage of economical reservation and transmission without burst or packet loss.

In a further embodiment of the invention, the arriving data packets are added to a stored burst. This has the advantage that only one burst has to be reserved instead of two successive bursts. Superior transmission and multiplexing gain is achieved, which is economical.

Additional advantages of the invention are mentioned in carrying out the invention.

An exemplary embodiment of the invention is explained by means of a drawing.

The drawing shows:

in figure 1 a ring network.

Figure 1 shows ten central nodes respectively ring nodes node_0, node_1, ..., node_9, so that each central node is connected with two other central nodes to form a ring network. These central nodes can be optical switches, network elements, routers or switches for an OBS or OPS network. Five border nodes Border_0, Border 1, ..., Border 4 are connected to a central node. Border node Border_0 is connected to central node node_0, border node Border_1 is connected to central node node_ 2, border node Border_2 is connected to central node node_4, border node Border_3 is connected to node central node_6 and the border node Border_4 is connected to the central node node_8.

The border nodes Border_0, Border 1, ..., Border 4 aggregate data packets from non-displayed sources to bursts, such as optical bursts, or transform them into optical packets as explained above. These bursts / packets are sent from the border node to the associated central node and beyond through the central network to a destination border node. The border nodes Border_0, Border 1, ..., Border 4 divide the bursts received by the OBS or OPS network respectively central nodes into data packets and send them to receivers or destinations not shown.

ES 2 306 830 T3

In such a ring network, the only nodes where blocking could take place are those that are connected to a border node. These are the core nodes node_0, node_2, node_4, node_6, and node_8. The rest of the core nodes, for example node_1, node_3, node_5, node_7 and node_9, simply forward the bursts or packets and are therefore free of blocking.

After sending a header of a burst or packet from a border node to an associated central node / ring node, two cases can occur.

First, the resources to transmit the burst / packet are available at the central node and a path is reserved according to the behavior already described. After a lag time the burst / packet at the border node is sent and the burst / packet is transmitted through the reserved paths to its destination.

Second, the resources for reserving a path and transmitting the burst / packet are not available. To avoid a collision, a response is sent back to the originating boundary node causing the burst to wait, until there is (one) additional transmission attempt (s).

After sending the header from the border node to the associated hub, the header is evaluated at the hub and the requested reservation is compared with previous reservations from other headers. In the event of a reservation conflict, a response is sent back to the originating border node. Said response is evaluated at the border node and the sending of the burst or packet is abandoned and the burst or packet is stored at the border node in the electrical domain.

The response comprises the time or occupancy time delay or a time or time delay for a future reservation or any other kind of message to transfer information from the border node to the central node.

Depending on the information sent to the border node and an implemented mechanism, this node may have a different behavior.

First, in the event of a reservation conflict, the central node determines a time or a time delay for an occupation of the central node and sends a response, in which said time or time delay is contained, to the border node of source. Said response is evaluated at the border node, sending the burst or packet is abandoned, the burst or packet is stored at the border node and after said time or time delay has elapsed, a header is sent back to the central node. , to try a new reservation for a transmission of said burst or packet.

Second, that in the event of a reservation conflict, the central node schedules a reservation for the burst or packet of the associated headend, determines a particular time or time delay for the transmission of said burst or packet and sends a response, in the one that is contained said time or particular time delay, to the border node of origin. There, said response is evaluated, the burst or packet is abandoned, the burst or packet is stored in the border node and after said time or time delay has elapsed, the burst is sent to the central node, to use the scheduled reservation. .

It is independent of which burst reservation mechanism is used, such as just-in-time (JIT, Just-In-Time), Horizon reservation mechanism (Horizon), or just enough time (JET, Just-Enough Time). If a burst / packet header from a border node reaches the central node, the node knows according to its reservation schedule until when it will be busy. This is due to old reservations by other headends received from neighboring central nodes. Therefore, the central node can deduce if the burst / optical packet associated with the headend will be blocked or not.

During burst storage at the border node the arriving data packets can be added to the stored burst. If this is done, a second reservation for a burst is avoided, which in the meantime contains incoming data packets. Therefore only one burst is sent, which is larger than before and has a higher multiplexing gain. Because one larger burst is sent instead of two bursts with corresponding headers, fewer resources are used.

Once a burst / packet is inserted into the ring, star or bus network, it will continue its way to the destination border node without any kind of blocking. If the network load increases, the optical bursts / packets that cannot be transported over the network will be stored at the border nodes. If the network load reaches a critical level, the buffers on the border nodes will eventually become overloaded and lose information. However, even under these conditions of high load the network (ring, star, bus) will not have losses.

At the border node the mentioned burst can be transformed and sent as an optical burst to a central node of a switched optical burst network. Also, at the border node the mentioned packet can be transformed and sent as an optical packet to a central node of an optical packet switched network.

The fundamental principles of the invention are:

nodes, where blocking could take place, send a response back to the corresponding border node if they foresee according to their burst reservation schedule that their burst or associated packet will be blocked.

ES 2 306 830 T3

If a border node receives such a response, the burst / packet is abandoned and the burst / packet is stored. In the case of an OBS network this could occur by storing an electrical version of the optical bursts / packets that have to be sent to the ring. Such border nodes are often an interface between the electrical and the optical domain.

If a border node sends a header and receives a reply back, it automatically knows that the associated burst / packet will not go through the next central node due to blocking. It is also possible that the response of a burst / packet that is about to be blocked carries some useful information from the central node on its way back to the border node. This information could be for example the exact moment when the central node is reserved for the burst / optical packet. Knowing this, the border node can immediately send the burst / optical packet just when the time / moment is reached.

The same principles described above by way of an example for a ring network can be applied analogously to a star network (a central node with many border nodes) and bus ("split or broken ring") without modifications.

The methods of the invention extend the principles of packet or burst switched networks, such as OBS and OPS networks, with a central node without burst or packet storage capabilities from the unconfirmed sending of a burst to a "back pressure" communication for the first hop from the border node to the central node in the case of conflict at the central node.

Additional advantages of the methods of the invention are:

• reduces the probability of network blocking to zero.

• the solution is particularly suitable for packet and burst switched networks.

• The solution is especially suitable for ring, star and bus networks.

• It is simple and easy. It does not require a lot of processing time either on the central nodes or on the border nodes.

• It does not consume bandwidth, since it does not require increasing the size of the optical headers.

• It is inexpensive since it does not require additional hardware. Furthermore, for OBS / OPS networks, storing information electrically at the border nodes is a low-cost solution compared to optical storage with fiber delay lines (FDL, Fiber Delay Lines).

• The procedure can be implemented in software.