Computational storage device, storage system including the same and operation method therefor
A computational storage device includes a control module, and a nonvolatile memory connected to the control module. The nonvolatile memory is configured to store a plurality of graph data elements, which comprises a plurality of nodes and a plurality of edges that connect at least a portion of the plurality of nodes to each other, in a first memory area and a second memory area each having a plurality of blocks and having different read speeds. The control module is configured to store a first graph data element of the plurality of graph data elements having a relatively high degree of association with one node of the plurality of nodes in the first memory area, and store a second graph data element of the plurality of graph data elements having a relatively low degree of association with the one node of the plurality of nodes in the second memory area.
1 . A computational storage device comprising:
a control module; and
a nonvolatile memory connected to the control module, and configured to store a plurality of graph data elements, which comprises a plurality of nodes and a plurality of edges that connect at least a portion of the plurality of nodes to each other, in a plurality of memory areas each having a plurality of blocks, and
wherein the control module is configured to:
train an artificial intelligence model based on the plurality of graph data elements; and
store a graph data element of the plurality of graph data elements in a designated block of the plurality of blocks based on training of the artificial intelligence model and an update frequency of the graph data element.
2 . The computational storage device of claim 1 , wherein the control module is further configured to:
store a graph data element of the plurality of graph data elements in at least one reservation page included in the designated block based on training the artificial intelligence model.
3 . The computational storage device of claim 1 , wherein the designated block is included in a one memory area of the plurality of memory areas.
4 . The computational storage device of claim 1 , wherein the artificial intelligence model is a GNN (Graph Neural Network) model.
5 . The computational storage device of claim 1 , wherein the control module is further configured to:
designate the designated block based on an update frequency of the graph data elements, the plurality of blocks, and/or a plurality of pages included in each of the plurality of blocks.
6 . The computational storage device of claim 1 , wherein the control module performs a garbage collection for the designated block when the designated block is full.
7 . The computational storage device of claim 6 , wherein the garbage collection is block-based operation.
8 . The computational storage device of claim 1 , wherein the designated block is designated in multiples.
9 . The computational storage device of claim 1 , wherein the graph data element is a first data graph element having a first update frequency, the control module stores a second graph data element of the plurality of graph data elements having a second update frequency in a remaining block of the plurality of blocks based on training of the artificial intelligence model, and the first update frequency is higher than the second update frequency.
10 . The computational storage device of claim 1 , wherein the designated block is designated through a host connected to the computational storage device.
11 . A method of operating a computational storage device, the method comprising:
training an artificial intelligence model based on a plurality of graph data elements comprising a plurality of nodes and a plurality of edges that connect at least a portion of the plurality of nodes to each other; and
storing a graph data element of the plurality of graph data elements in a designated block of a plurality of blocks based on training of the artificial intelligence model and an update frequency of the graph data element.
12 . The method of claim 11 , further comprising:
storing a graph data element of the plurality of graph data elements in at least one reservation page included in the designated block based on training the artificial intelligence model.
13 . The method of claim 11 , further comprising:
designating the designated block based on an update frequency of the graph data elements, the plurality of blocks, and/or a plurality of pages included in each of the plurality of blocks.
14 . The method of claim 11 , further comprising:
performing a garbage collection for the designated block when the designated block is full.
15 . The method of claim 14 , wherein the garbage collection is block-based operation.
16 . The method of claim 11 , wherein the designated block is designated in multiples.
17 . The method of claim 11 , wherein the graph data element is a first data graph element having a first update frequency, and the method further comprising:
storing a second graph data element of the plurality of graph data elements having a second update frequency in a remaining block of the plurality of blocks based on training of the artificial intelligence model, and the first update frequency is higher than the second update frequency.
18 . The method of claim 11 , wherein the designated block is designated through a host connected to the computational storage device.
19 . A network system comprising:
a storage system that performs one or more graph processing operations; and
a cloud server connected to the storage system and transmitting and receiving data related to the one or more graph processing operations,
wherein the one or more graph processing operations comprise:
training an artificial intelligence model based on a plurality of graph data elements comprising a plurality of nodes and a plurality of edges that connect at least a portion of the plurality of nodes to each other, and
storing a graph data element of the plurality of graph data elements in a designated block of a plurality of blocks based on training of the artificial intelligence model and an update frequency of the graph data element.
20 . The network system of claim 19 , wherein the artificial intelligence model is a GNN (Graph Neural Network) model.