Cluster system, load distribution method, optimization client program, and arbitration server program

• US 7,623,460 B2
• Filed: 04/20/2007
• Issued: 11/24/2009
• Est. Priority Date: 05/19/2006
• Status: Active Grant

• Current Assignee: NEC Corporation
• Original Assignee: NEC Corporation
• Inventors: Miyazaki, Manabu
• Primary Examiner(s): PHAM, BRENDA H
• Application Number:
  US11/737,833
• Publication Number:
  US 20070286184A1
• Time in Patent Office: 949 Days
• Field of Search: 370/229, 370/232, 370/233, 370/234, 370/235, 370/237, 370/235.1, 370/238, 370/242, 370/248, 370/252, 370/253, 370/254
• US Class Current: 370/235
• CPC Class Codes: H04L 47/10 Flow control; Congestion control
• CPC Class Codes: H04L 47/11 Identifying congestion
• CPC Class Codes: H04L 47/125 by balancing the load, e.g. traffic engineering

Abstract

An exemplary cluster system according to the present invention includes a first node including a plurality of paths respectively connected to I/O slots via a switch and a second node including a plurality of paths connected to the switch, wherein the first node includes a first load measuring part which measures a first load for each path thereof, and the switch makes switching to disconnect an I/O slot from the first node and connect the I/O slot to the second node when the first load of the path connected to the I/O slot is higher than a first threshold.

Claims

1. A cluster system comprising:

a first node including a plurality of paths respectively connected to I/O slots via a switch and;

a second node including a plurality of paths connected to said switch, wherein

said first node includes a first load measuring part which measures a first load for each path thereof, and

said switch makes switching to disconnect an I/O slot from said first node and connect said I/O slot to said second node when said first load of the path connected to said I/O slot is higher than a first threshold.

2. The cluster system according to claim 1, wherein

said first node further includes a second load measuring part which measures a second load which is a total load of all of said paths thereof, and

said witching is made when said second load is higher than a second threshold.

3. The cluster system according to claim 2, wherein said second node includes a first load measuring part which measures a first load for each path thereof and one of said plurality of paths is determined to have a lowest load in said cluster system based on said first load.

4. The cluster system according to claim 3, further comprising:

an arbitration node, wherein

each of said first node and said second node sends said first load measured value to said arbitration node,

said arbitration node issues a configuration change request based on said first load measured values received, and

each of said first node and said second node controls said switch to make said switching.

5. The cluster system according to claim 4, further comprising:

a switch management network for connecting said arbitration node, said first node, said second node and said switch, wherein

each of said first node and said second node sends said first load measured value to said arbitration node and controls said switch through said switch management network.

6. The cluster system according to claim 5, wherein

said arbitration node generates configuration information which specifies a task to be processed by each of said first node and said second node based on said first load measured values received and sends said configuration information to each of said first node and second node.

7. The cluster system according to claim 6, wherein

each of said first node and said second node further includes performance statistical information that records said first load measured value in a time series and a user policy determining a condition to change a connection between an I/O slot and, said first node or said second node,

each of said first node and said second node sends said performance statistical information to said arbitration node, and

said arbitration node issues said configuration change request generated based on said performance statistical information and said user policy.

8. The cluster system according to claim 7, wherein

said first node further includes a processor, a PCI bridge and a northbridge connected between said processor and said PCI bridge;

said plurality of paths are a plurality of PCI buses connected to said PCI bridge,

said PCI bridge includes said first load measuring part and said northbridge includes a second load measuring part, and

said processor issues said arbitration request to said arbitration node.