Graph processing system
A graph comprising nodes and edges is stored by a distributed system as a collection of nodes and half-edges stored with their respective nodes. A message processor is associated with a node as needed to process messages passed between nodes, such that a given node has zero or one message processor assigned to it at a given time. Queries of the graph are resolved by processing a first portion of the query at a first node, and forwarding the results with the remaining portions of the query to a node linked by an edge to the present node.
1 . A system, comprising:
one or more processors and at least one memory comprising instructions that, in response to execution by the one or more processors, cause the system to at least:
store a first node in a graph, the first node linked to a second node in the graph by a first half-edge associated with the first node, and a second half-edge associated with the second node;
cause a message processor to process one or more messages directed to the first node, wherein the message processor causes information indicative of properties of the first node to be stored in one or more records associated with records of the first half-edge and the second half-edge; and
generate one or more responses to one or more queries of the graph based, at least in part, on identification of one or more patterns in the graph, the one or more patterns comprising the first node and the second node.
2 . The system of claim 1 , wherein at least one portion of the one or more queries is lazily evaluated.
3 . The system of claim 1 , wherein one or more additional processors on one or more additional systems store additional records indicative of additional nodes of the graph.
4 . The system of claim 1 , wherein the first half-edge and the second half-edge collectively define a link from the first node to the second node.
5 . The system of claim 1 , wherein the one or more responses to the one or more queries are generated based, at least in part, on one or more messages processed by the message processor up to a defined point.
6 . The system of claim 1 , wherein processing of messages is constrained to a subset of nodes of the graph.
7 . The system of claim 1 , wherein the message processor is assigned to process messages directed to the first node.
8 . A computer-implemented method, comprising:
storing a first node in a graph, the first node linked to a second node in the graph by a first half-edge associated with the first node, and a second half-edge associated with the second node;
causing a message processor to process one or more messages by storing one or more records associated with records of the first half-edge and the second half-edge; and
generating one or more responses to one or more queries based, at least in part, on identification of one or more patterns in the graph, the one or more patterns comprising the first node and the second node.
9 . The method of claim 8 , wherein one or more nodes of the graph are associated with one or more globally unique identifiers.
10 . The method of claim 8 , wherein nodes of the graph are partitioned between a plurality of computing devices.
11 . The method of claim 8 , wherein a first computing device stores the first node and the first half-edge, and a second computing device stores the second node and the second half-edge.
12 . The method of claim 8 , further comprising storing a property associated with the link between the first and second nodes by at least storing an additional node comprising the property associated with the link, and associating the additional node with the first half-edge and the second half-edge.
13 . The method of claim 8 ,
wherein one or more portions of the one or more queries are lazily evaluated.
14 . The method of claim 8 , wherein the one or more queries are evaluated with respect to a historical state.
15 . The method of claim 8 , wherein the one or more queries are standing queries.
16 . A non-transitory computer-readable storage medium comprising instructions that, when executed by at least one processor of at least one computing device, cause the at least one computing device to:
store a first node in a graph, the first node linked to a second node in the graph by a first half-edge associated with the first node, and a second half-edge associated with the second node;
cause a message processor to process one or more messages by at least storing one or more records associated with records of the first half-edge and the second half-edge; and
generate one or more responses to one or more queries based, at least in part, on identification of one or more patterns in the graph, the one or more patterns comprising the first node and the second node.
17 . The non-transitory computer-readable storage medium of claim 16 , wherein nodes of the graph are partitioned between a plurality of computing devices.
18 . The non-transitory computer-readable storage medium of claim 16 , wherein the one or more responses to the one or more queries are generated based, at least in part, on processing a first portion of the one or more queries applicable to the first node and forwarding a second portion of the one or more queries to one or more other nodes.
19 . The non-transitory computer-readable storage medium of claim 16 , wherein the one or more responses to the one or more queries are generated based, at least in part, on one or more messages processed by the message processor up to a defined point in time.
20 . The non-transitory computer-readable storage medium of claim 16 , wherein at least one portion of the one or more queries is lazily evaluated.