On-demand instantiation in a high-performance computing (HPC) system

• US 20060117208A1
• Filed: 11/17/2004
• Published: 06/01/2006
• Est. Priority Date: 11/17/2004
• Status: Active Grant

• Current Assignee: Raytheon Company (Rtx Corporation)
• Original Assignee: Raytheon Company (Rtx Corporation)
• Inventors: Davidson, Shannon V.
• Granted Patent: US 8,244,882 B2
• Time in Patent Office: Days
• US Class Current: 714/4
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1006 with static server selection, e.g. the same server being selected for a specific client
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/1023 based on a hash applied to IP addresses or costs

Abstract

In one embodiment, a method for on-demand instantiation in a high-performance computing (HPC) system includes receiving a connection request from a client specifying a first port number and a first host name advertised externally with respect to an HPC server including a cluster of nodes, identifying a service at the HPC server corresponding to the first port number and the first host name, determining whether the identified service is available, and, if the identified service is available, instantiating a host providing the identified service at one or more nodes in the cluster.

Claims

1. Logic for on-demand instantiation in a high-performance computing (HPC) system, the logic encoded in a computer-readable medium and when executed operable to:

receive a connection request from a client specifying a first port number and a first host name, the first port number and the first host name advertised externally with respect to an HPC server comprising a cluster of nodes;

identify a service at the HPC server corresponding to the first port number and the first host name;

determine whether the identified service is available; and

if the identified service is available, instantiate a host providing the identified service at one or more nodes in the cluster.

2. The logic of claim 1, wherein the connection request is a Transmission Control Protocol (TCP) connection request or a User Datagram Protocol (UDP) connection request.

3. The logic of claim 1, further operable to:

access a list of services at the HPC server comprising a plurality of entries that each specify;

a service; and

a port number and a host name advertised externally with respect to the HPC server that correspond to the service; and

identify, according to the list of services at the HPC server, the service at the HPC server corresponding to the first port number and the first host name.

4. The logic of claim 1, further operable to:

access a list of services at the HPC server comprising a plurality of entries that each specify;

a service; and

a port number and a host name internal to the HPC server that correspond to a host that, when executed at one or more nodes in the cluster, provides the service; and

instantiate, according to the list of services at the HPC server, the host providing the identified service at one or more nodes in the cluster.

5. The logic of claim 1, further operable to:

access a list of services at the HPC server comprising a plurality of entries that each specify a service and one or more rules indicating whether the service is available; and

determine, according to the list of services, whether the identified service is available.

6. The logic of claim 5, wherein the rules indicate whether the service is available based on a time associated with the connection request from the client.

7. The logic of claim 1, further operable, if the identified service is unavailable, to block a port on the HPC server corresponding to the first port number to prevent the client from accessing the identified service.

8. The logic of claim 1, further operable to:

access a boot image, a file system, and an operating system (OS) configuration file at the HPC server corresponding to the host providing the identified service at one or more nodes in the cluster; and

use the boot image, the file system and the OS configuration file to instantiate the host providing the identified service at one or more nodes in the cluster.

9. The logic of claim 1, wherein the identified service is a web server.

10. The logic of claim 1, further operable to spoof an address of the host providing the identified service at one or more nodes in the cluster and initiate a connection sequence on behalf of the host while the host is booting.

12. The logic of claim 1, wherein a node in the cluster of nodes is a central processing unit (CPU) coupled to two switches.

27. The logic of claim 1, wherein:

the address is an Internet Protocol over Media Access Control (IP/MAC) address; and

the connection sequence is a Transmission Control Protocol over IP (TCP/IP) connection sequence or a User Datagram Protocol (UDP) over IP (UDP/IP) connection sequence.

28. The logic of claim 1, wherein the cluster of nodes comprises a plurality of nodes coupled to each other according to a topology comprising a three dimensional torus.

11. (canceled)

11-1. (canceled)

13. A method for on-demand instantiation in a high-performance computing (HPC) system, the method comprising:

receiving a connection request from a client specifying a first port number and a first host name, the first port number and the first host name advertised externally with respect to an HPC server comprising a cluster of nodes;

identifying a service at the HPC server corresponding to the first port number and the first host name;

determining whether the identified service is available; and

if the identified service is available, instantiating a host providing the identified service at one or more nodes in the cluster.

14. The method of claim 13, wherein the connection request is a Transmission Control Protocol (TCP) connection request or a User Datagram Protocol (UDP) connection request.

15. The method of claim 13, further comprising:

accessing a list of services at the HPC server comprising a plurality of entries that each specify;

a service; and

a port number and a host name advertised externally with respect to the HPC server that correspond to the service; and

identifying, according to the list of services at the HPC server, the service at the HPC server corresponding to the first port number and the first host name.

16. The method of claim 13, further comprising:

accessing a list of services at the HPC server comprising a plurality of entries that each specify;

a service; and

a port number and a host name internal to the HPC server that correspond to a host that, when executed at one or more nodes in the cluster, provides the service; and

instantiating, according to the list of services at the HPC server, the host providing the identified service at one or more nodes in the cluster.

17. The method of claim 13, further comprising:

accessing a list of services at the HPC server comprising a plurality of entries that each specify a service and one or more rules indicating whether the service is available; and

determining, according to the list of services, whether the identified service is available.

18. The method of claim 17, wherein the rules indicate whether the service is available based on a time associated with the connection request from the client.

19. The method of claim 13, further comprising, if the identified service is unavailable, blocking a port on the HPC server corresponding to the first port number to prevent the client from accessing the identified service.

20. The method of claim 13, further comprising:

accessing a boot image, a file system, and an operating system (OS) configuration file at the HPC server corresponding to the host providing the identified service at one or more nodes in the cluster; and

using the boot image, the file system and the OS configuration file to instantiate the host providing the identified service at one or more nodes in the cluster.

21. The method of claim 13, wherein the identified service is a web server.

22. The method of claim 13, further comprising spoofing an address of the host providing the identified service at one or more nodes in the cluster and initiating a connection sequence on behalf of the host while the host is booting.

23. The method of claim 13, wherein:

the address is an Internet Protocol over Media Access Control (IP/MAC) address; and

the connection sequence is a Transmission Control Protocol over IP (TCP/IP) connection sequence or a User Datagram Protocol (UDP) over IP (UDP/IP) connection sequence.

24. The method of claim 13, wherein the cluster of nodes comprises a plurality of nodes coupled to each other according to a topology comprising a three dimensional torus.

25. The method of claim 13, wherein a node in the cluster of nodes is a central processing unit (CPU) coupled to two switches.

26. A system for on-demand instantiation in a high-performance computing (HPC) system, the system for on-demand instantiation in an HPC system comprising:

means for receiving a connection request from a client specifying a first port number and a first host name, the first port number and the first host name advertised externally with respect to an HPC server comprising a cluster of nodes;

means for identifying a service at the HPC server corresponding to the first port number and the first host name;

means for determining whether the identified service is available; and

means for, if the identified service is available, instantiating a host providing the identified service at one or more nodes in the cluster.