Loosely coupled system environment designed to handle a non-disruptive host connection switch after detection of an error condition or during a host outage or failure

• US 5,802,258 A
• Filed: 05/03/1996
• Issued: 09/01/1998
• Est. Priority Date: 05/03/1996
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Chen, Shawfu
• Primary Examiner(s): Beausoliel, Jr., Robert W.
• Assistant Examiner(s): HUA, LY
• Application Number:
  US08/642,715
• Time in Patent Office: 851 Days
• Field of Search: 395/182.08, 395/200.03, 395/200.07, 395/200.12, 395/182.02
• US Class Current: 714/10
• CPC Class Codes: G06F 11/1443 Transmit or communication errors

Abstract

A system for maintaining a non-disruptive connection between a local host and a remote host is introduced that allows the re-establishment of processing communications between the remote host and a new host in case the first host originally in processing communication with the remote host experiences an outage. The present invention particularly relates to loosely coupled environments connecting to a TCP/IP network. The system uses a communication box or an offload frontend box for holding all socket associations between the remote host and the local hosts. Each socket connection associates the remote host to an application being run on the local host but maintained by the communication or the offload frontend box. The offload frontend box detects a host outage through its channel connectivity and holds all processing and communication data related to the failing host until a new connection is established between a second host and the remote host, at which time the box transfers all such data to the second replacing host.

Claims

1. In a loosely coupled computer environment, a system for providing non-disruptive connection switch from a first host to a second host during said first host'"'"'s outage, said system comprising:

a plurality of hosts and a plurality of remote hosts, with at least one of said remote hosts being in processing communication with at least a first host through the use of a communication box;

said communication box maintaining all information about all said hosts and said remote hosts connected to said system;

said communication box having error detection and recovery mechanism, as to allow said communication box to detect an error condition associated with said first host in processing communication with one of said remote host;

said communication box having a storage location for keeping all processing data and communication information between said first host and said remote host in which it was in communication with during first host'"'"'s outage until a second replacing host re-establishes processing communication through said communication box with said remote host.

2. The system of claim 1, wherein said communication box is an offload frontend box able to conduct processing communication protocol for the environment in order to allow said loosely coupled environment to conduct computation and other data server functions.

3. The system of claim 2, wherein said communication protocol is a TCP/IP protocol.

4. The system of claim 2, further comprising a plurality of channels with at least one channel being assigned to each of said hosts in said system for communication between said offload frontend box and said hosts, said offload frontend box detecting said host outages through said channels.

5. The system of claim 2, said offload frontend box having a storage location for maintaining applications for all said hosts as requested by any of said remote hosts.

6. The system of claim 5, wherein said offload frontend box has at least one socket and that said offload frontend box associates said socket to an application in said hosts, said application being maintained by said offload frontend box.

7. The system of claim 6, wherein said offload frontend box has a recovery mechanism allowing a second replacing host to re-establish an association with said remote host by taking over said association originally assigned to said first host, said offload frontend box then forwarding all processing and communication data to said remote host after said association has been taken over.

8. The system of claim 7, wherein said recovery mechanism further comprises an acknowledgment element that needs to be set before any further communication or processing data can be forwarded to said second host;

said acknowledgement element being set through the receipt of a signal from said second host signifying that said second host has successfully accepted said socket assigned to said first host and processing communication has been re-established with said remote host.

9. The system of claim 1, wherein a plurality of hosts are in processing communication with a plurality of remote hosts, and at least two hosts experience an outage simultaneously.

10. The system of claim 1, wherein said communication box is an offload frontend box performing several functions.

11. The system of claim 10, wherein said functions comprise flow control, error detection and recovery and routing of message between said remote hosts and said hosts.

12. A system for maintaining a non-disruptive connection between at least one host and a remote host in a loosely coupled computer environment, said system comprising:

a plurality of hosts and a plurality of remote hosts, with at least one of said remote hosts being in processing communication with a first host through the use of a communication box;

a plurality of channel connections for connecting said communication box with said hosts;

an error detection mechanism provided in said communication box as to allow detection of a host outage through said channel connectivity between said hosts and said communication box;

said connection box having a plurality of socket connections to associate each host to an application maintained by said communication box, said applications being run by said hosts in processing communication;

a recovery mechanism provided in said communication box, said recovery mechanism having storage areas for maintaining all communication and processing data when a first processing host takes an outage until a second host re-establishes processing communication with said remote host;

said recovery mechanism further having a switching element for reassigning said socket connection for said first processing host to a second processing host upon said first host'"'"'s outage.

13. The system of claim 12, wherein said recovery mechanism further comprises an acknowledgment element that needs to be set before any further communication or processing data can be forwarded to said second host;

said acknowledgement element being set through the receipt of a signal from said second host signifying that said second host has successfully accepted said socket assigned to said first host and processing communication has been re-established with said remote host.

14. The system of claim 13, wherein said communication box is an offload frontend box for processing communication protocol in order to allow said loosely coupled environment to conduct computation and other data server functions.

15. The system of claim 14, wherein said communication protocol is TCP/IP protocol.

16. The system of claim 13, wherein a plurality of hosts are in processing communication with a plurality of remote hosts, and at least two hosts experience an error condition simultaneously.

17. The system of claim 12, wherein said communication box is an offload frontend box performing several functions.

18. The system of claim 17, wherein said functions comprise flow control, error detection and recovery and routing of messages between said remote hosts and said hosts.

19. A system for maintaining a non-disruptive connection between at least one host and a remote host in a loosely coupled computer environment, said system comprising:

a plurality of hosts and a plurality of remote hosts, with at least one of said remote hosts being in processing communication with a first host through the use of an offload frontend box;

said offload frontend box having communication means for processing TCP/IP protocol in order to allow said loosely coupled environment to conduct computation and other data server functions;

a plurality of channel connections for connecting said offload frontend box with said hosts;

an error detection mechanism provided in said offload frontend box as to allow detection of a host outage through said channel connections between said hosts and said offload frontend box;

said offload frontend box having a plurality of socket connections to associate each host to an application maintained by said communication box but assigned to said any processing hosts;

a recovery mechanism provided in said offload frontend box, said recovery mechanism having storage areas for maintaining all communication and processing data when a first processing host takes an outage until a second host re-establishes processing communication with said remote host;

said recovery mechanism further having a switching element for reassigning said socket connection for said first processing host to a second processing host upon said first host'"'"'s outage;

said recovery mechanism further comprising an acknowledgment element that needs to be set before any further communication or processing data can be forwarded to said second host;

said acknowledgement element being set through the receipt of a signal from said second host signifying that said second host has successfully accepted any socket assigned to said first host and processing communication has been re-established with said remote host.

20. The system of claim 19, wherein a plurality of remote hosts are in communication with a plurality of hosts and an error is detected simultaneously with at least two hosts, with a plurality of other hosts being available for reestablishing communication and processing relation with said plurality of remote hosts.

21. The system of claim 19, wherein said communication box is an offload frontend box performing several functions.

22. The system of claim 21, wherein functions comprise flow control, error detection and recovery and routing of messages between said remote hosts and said hosts.