Minimizing micro-interruptions in high-performance computing
Data storage systems and methods for storing data in computing nodes of a super computer or compute cluster are described herein. The super computer storage may be coupled with a primary storage system. In addition to a CPU and memory, non-volatile memory is included with the computing nodes as local storage. The super computer includes a plurality of computing groups, each including a plurality of computing nodes. There is one burst buffer fabric per group and one input/output node per group. When data bursts occur, data may be stored by a first computing node on the local storage of a second computing node in the computing group through the burst buffer fabric without interrupting the CPU in the second computing node. Further, the local storage of other computing nodes may be used to store redundant copies of data from a first computing node to make the super computer data resilient.
1. A data storage method comprising:
a CPU of a first computing node of a super computer issuing a data write request;
evaluating an availability of local storage in the first computing node;
evaluating an availability of local storage in at least one other computing node of a plurality of computing nodes configured as a computing group in which the first computing node is a member, including querying an input/output node for the computing group to obtain an identifier of the at least one other computing node in the same computing group with available local storage;
evaluating storage policies in view of the evaluating the availability of local storage in the first computing node and the evaluating the availability of local storage in the at least one other computing node in the computing group;
writing data to local storage of at least one of the other computing nodes in the computing group according to the storage policies and the availability of local storage both in the first computing node and in other computing nodes, including writing data from the first computing node to the local storage of the at least one other computing node through a burst buffer fabric such that a CPU in the at least one other computing node is not interrupted, wherein the burst buffer fabric conforms to a storage device access standard.
2. The data storage method of claim 1 wherein the local storage is non-volatile memory.
3. The data storage method of claim 2 wherein the non-volatile memory is silicon-based.
4. The data storage method of claim 1 wherein the storage device access standard is one of a serial attached small computer system interface (SAS) standard or a peripheral component interconnect express (PCI-EXPRESS) standard.
5. The data storage method of claim 1 wherein the evaluating the querying includes requesting an address of local storage in another computing node in the computing group having a specified amount of storage space available.
6. The data storage method of claim 1 wherein the evaluating the availability of local storage in the first computing node comprising checking whether the local storage is accessible and/or full.
7. The data storage method of claim 1 further comprising:
reporting location information of data written to local storage in the at least one other computing node to the input/output node.
8. The data storage method of claim 7 wherein the reporting the location information to the input/output node is achieved over one of a system fabric and the burst buffer fabric.
9. A data retrieval method comprising:
a first computing node of a plurality of computing nodes of a computing group checking whether needed data is in its local storage;
when the data is not available from the local storage of the first computing node, the first computing node requesting the data from an I/O node for the computing group;
when the data is not in primary storage, the first computing node receiving from the I/O node a location of a second computing node of the computing group where the needed data is located;
the first computing node requesting the needed data through a burst buffer fabric from the second computing node, wherein the burst buffer fabric connects all of the computing nodes in the computing group with each other and with the I/O node of the computing group, and wherein the burst buffer fabric conforms to a storage device access standard;
the first computing node receiving the requested data from the second computing node over the burst buffer fabric such that a CPU in the second computing node is not interrupted.
10. The data retrieval method of claim 9 wherein checking whether needed data is in local storage includes checking whether the local storage is accessible.
11. The data retrieval method of claim 9 wherein the local storage is non-volatile memory.
12. The data retrieval method of claim 9 wherein the non-volatile memory is silicon-based.
13. The data retrieval method of claim 9 wherein the storage device access standard is one of a serial attached small computer system interface (SAS) standard or a peripheral component interconnect express (PCI-EXPRESS) standard.
14. The data retrieval method of claim 9 further comprising:
when the data is in the primary storage, the I/O node
requesting the requested data from an appropriate primary storage location,
receiving the requested data to from a storage server,
providing the requested data to the requesting computing node.
15. The data retrieval method of claim 14 wherein the providing the requested data to the requesting computing node is achieved over a system fabric.
16. A super computer comprising:
a plurality of computing groups, each computing group having a plurality of computing nodes, each computing node including a CPU, a memory, a local storage, and a communications unit;
a plurality of input/output nodes corresponding to a number of computing groups;
a system fabric over which the computing nodes communicate with each other and with the input/output nodes, wherein the system fabric is a high speed interconnect, such that there is one system fabric for the super computer;
a plurality of burst buffer fabrics corresponding to the number of computing groups over which the computing nodes in a particular computing group communicate with each other and with the input/output node for the computing group, wherein the burst buffer fabric conforms to a storage device access standard,
wherein each of the plurality of computing nodes in one computing group are configured to access the local storage of other computing nodes in the one computing group via the burst buffer fabric for the one computing group without interrupting the CPUs in the other computing nodes in the one computing group.
17. The super computer of claim 16 wherein the local storage is non-volatile memory.
18. The super computer of claim 16 wherein the non-volatile memory is silicon-based.
19. The super computer of claim 16 wherein the storage device access standard is one of a serial attached small computer system interface (SAS) standard or a peripheral component interconnect express (PCI-EXPRESS) standard.
20. The super computer of claim 16 wherein the memory is random access memory (RAM).
21. The super computer of claim 16 wherein the high speed interconnect conforms to one of INFINIBAND, CASCADE, GEMINI standards.
22. The super computer of claim 16 wherein the high speed interconnect is one of wired or fiber optic.
23. The super computer of claim 16 wherein the input/output node is a server computer.
24. The super computer of claim 16 wherein the input/output nodes each include a database to store location information for stored data.