Adaptive private network asynchronous distributed shared memory services
A highly predicable quality shared distributed memory process is achieved using less than predicable public and private internet protocol networks as the means for communications within the processing interconnect. An adaptive private network (APN) service provides the ability for the distributed memory process to communicate data via an APN conduit service, to use high throughput paths by bandwidth allocation to higher quality paths avoiding lower quality paths, to deliver reliability via fast retransmissions on single packet loss detection, to deliver reliability and timely communication through redundancy transmissions via duplicate transmissions on high a best path and on a most independent path from the best path, to lower latency via high resolution clock synchronized path monitoring and high latency path avoidance, to monitor packet loss and provide loss prone path avoidance, and to avoid congestion by use of high resolution clock synchronized enabled congestion monitoring and avoidance.
1. A method for using distributed shared memory for communicating between loosely coupled computing systems, the method comprising:
calibrating a first time in a first node and a second time in a second node based on a current time received in the first node and in the second node from a node having a calibrated shared reference clock;
transmitting a block of memory from the first node over a first communication path to the second node and a duplicate block of memory from the first node over a second communication path, wherein the second communication path is different than the first communication path and wherein the block of memory and the duplicate block of memory are both tagged in the first node with a corresponding send time according to the calibrated first time;
selecting the block of memory or the duplicate block of memory as the first block of memory received in the second node and tagged with an arrival time according to the calibrated second time, wherein the send time and arrival time of the first block of memory received in the second node and the send time and arrival time of the duplicate block of memory received in the second node are analyzed taking into account previous transmission measurements at the second node and used to generate a latency report for the first communication path and the second communication path that is recorded in a second node path quality report; and
transmitting the second node path quality report from the second node to the first node, wherein path selection for transmission of a second block of memory from the first node is based on the generated latency report.
2. The method of claim 1 , wherein the calibrating is repeated over time to adjust the first time in the first node and the second time in the second node to the current time in the node having the calibrated shared reference clock.
3. The method of claim 1 , wherein the calibrating is in response to a first clock request message from the first node and in response to a second clock request message from the second node.
4. The method of claim 1 , wherein the calibrating comprises:
transmitting a clock request from the first node over a first path to the node having the calibrated shared reference clock and a duplicated clock request from the first node over a second path to the node having the calibrated shared reference clock; and
selecting the clock request or the duplicated clock request as the first received clock request in the node having the calibrated shared reference clock which responds to the first node with the current time over the first path or the second path corresponding to the path with the lowest latency associated with the first received clock request.
5. The method of claim 4 further comprises:
transmitting the current time from the node having the calibrated shared reference clock over a third path to the first node;
transmitting a current time of a second transmit from the node having the calibrated shared reference clock over a fourth path to the first node; and
selecting the current time or the current time of the second transmit as the first received current time in the first node, wherein the round trip latency of the clock request is minimized.
6. The method of claim 1 further comprising:
transmitting a first node path quality report from the first node to the second node, wherein path selection for transmission of the second node path quality report is based on the first node path quality report.
7. The method of claim 1 , wherein congestion is determined in the second node by comparing changes in a difference between send time and receive time of packets transmitted from the first node, wherein congestion is reduced with a current difference that is reduced from a previous measured difference and congestion is increased with a current difference that is greater than a previous delta.