Classical implementation of entanglement
Computational nodes are connected to each other via multiple point to point communication ports to form a mesh network. When a node receives such routing information, it will be a set of {Route ID, Forwarding Ports} pair, where the ports can be multiple targets including the node itself. When all the nodes within the routing path receive such information, the mesh system can handle the packet transfer for the given Route ID. Each node looks up the table with Route ID and determines where to forward the packet.
1. A computer-implemented method for controlling a packet transfer in a data communication network, where nodes are connected to each other with a point-to-point communication link for a packet to reach from a source node to a destination node comprising the steps of:
establishing a self-generating continually circulating packet of shared information between at least two nodes, the arrival of which creates an event that drive a multi-state machine at each end of the communication link;
allocating a unique table entry for given logical routing address on each node;
exchanging a table entry memory address offset with a neighbor node within a routing path in the communication network;
holding the neighbor node entry offset in a corresponding table entry; and
transferring the packet according to the logical address, wherein upon transferring the packet the source node finds the routing table entry associated with a logical address and update a specific field on the packet with the table offset value for the neighbor node, wherein the node on the routing path uses the field on a received packet to lookup its own table entry, updates the field with a neighbor value on the packet, and forwards the packet to the destination node.
2. The computer-implemented method of claim 1 , wherein the method manages conserved and equal quantities.
3. The computer-implemented method of claim 1 , wherein the method performs exactly-once semantics.
4. The computer-implemented method of claim 1 , wherein the method enables credit-based control flow.
5. The computer-implemented method of claim 1 , wherein the method enables load balancing.
6. The computer-implemented method of claim 1 , wherein the table entry has a bit vector to indicate multiple ports for a given logical address, and a bit to indicate a packet destination on the node itself.
7. The computer-implemented method of claim 1 , wherein the method further comprises the steps of:
receiving a packet by a node;
forwarding the packet to all indicated ports; and
forwarding the packet to a node upper computation layer if indicated as the destination to provide multicast capability with a single logical address.
8. The computer-implemented method of claim 1 , wherein the entry has at least one forwarding port with the entry offset for the next node, and one port entry for the backward path with the entry offset for the backward node.
9. The computer-implemented method of claim 1 , wherein the entry has only one forwarding port and one entry offset of the corresponding node to minimize the size of one entry.
10. The computer-implemented method of claim 4 , wherein in the credit-based control flow, two entries are bundled to represent a bidirectional data transfer mechanism.
11. The computer-implemented method of claim 4 , wherein multiple entries are bundled to represent a multicast data transfer mechanism.