High availability on-line transaction processing system

• US 5,996,001 A
• Filed: 02/10/1997
• Issued: 11/30/1999
• Est. Priority Date: 09/27/1994
• Status: Expired due to Fees

• Current Assignee: Philip Quarles, Stephanie Quarles
• Original Assignee: Philip Quarles, Stephanie Quarles
• Inventors: Quarles, Stephanie, Quarles, Philip
• Primary Examiner(s): An, Meng-Ai T.
• Assistant Examiner(s): Nguyen, Dzung C
• Application Number:
  US08/799,644
• Time in Patent Office: 1,023 Days
• Field of Search: 395/200.3, 395/200.53, 395/200.57, 395/200.79, 395/200.81, 395/200.33, 709/200, 709/223, 709/249, 709/101, 709/203, 709/208, 714/4, 714/43, 714/56, 714/712, 714/717
• US Class Current: 709/203
• CPC Class Codes: G06F 11/2023 Failover techniques
• CPC Class Codes: G06F 11/2038 with a single idle spare processing component
• CPC Class Codes: G06F 11/2048 where the redundant components share neither address space nor persistent storage
• CPC Class Codes: G06F 11/2097 maintaining the standby controller/processing unit updated initialisation or re-synchronisation thereof G06F11/1658 and subgroups

Abstract

A high availability on-line transaction processing (OLTP) system utilizes both hardware and software. Geographically separated primary (live) and backup (shadow) communications servers link live and shadow OLTP systems to geographically separated live and shadow Wide Area Networks and remote client computers. The remote client computers communicate with the live and shadow OLTP systems through their respective live and shadow WANs and communications servers. The live OLTP system sends "keep-alive" messages to the shadow system via the dedicated circuits on a frequent basis. If the shadow OLTP system does not receive a keep-alive message from the live system within a designated time interval, it sends "probe messages" to the live communications servers and live OLTP system, via the shadow WAN, client computers and live WAN to determine their status. If the live communications servers respond to the probe message(s) but the live OLTP system does not, the shadow OLTP system deduces that the live OLTP system failed and automatically takes over the role of the live system without operator intervention.

Claims

1. A method for maintaining a high availability OLTP system having a live and shadow OLTP system including live and shadow WANs, live and shadow communications servers and remote client computers, said method comprising the steps of:

a) establishing a connection between the live and shadow OLTP systems;

b) establishing a direct connection between the client computers and their respective live and shadow communications servers;

c) appending client computer identification data to the input messages the client computers send to the live and shadow OLTP systems; and

d) sending probe messages from the shadow OLTP to the live communications servers and live OLTP system via a shadow wide area network (WAN) and the client computers wherein this step comprises the steps of;

i) maintaining a dynamic connection list in the shadow OLTP system that contains the network addresses of every live communications server that is in session with a client computer and the network addresses of those client computers;

ii) sending a probe message from the shadow OLTP system to the respective shadow communications servers;

iii) relaying a probe message from the shadow communications server to the respective client computer;

iv) relaying the probe messages from the client computer to its live communications server;

v) returning a live communications server probe response message from the live communications server to the client computer;

vi) relaying the probe message from the live communications server to the live OLTP system;

vii) returning a probe response message from the live OLTP system to the live communications server;

viii) relaying the live OLTP probe response message from the live communications server to the client computer;

ix) relaying the live communications server probe response message from the client computer to the shadow communications server;

x) relaying the live OLTP probe response message from the client computer to the shadow communications server;

xi) relaying the probe response message from the shadow communications server to the shadow OLTP systems;

xii) monitoring the probe response messages from the live communications servers and live OLTP system, said step of monitoring performed by the shadow OLTP system;

xiii) determining the live OLTP system failed when probe response messages are received by the shadow OLTP system from the live communications servers but not from the live OLTP system;

xiv) placing the shadow OLTP system on line to take over the role of the live OLTP system without operator intervention when the shadow OLTP system determines the live OLTP system has failed;

xv) determining whether there are client computers in current session with live communications servers and having the shadow system send probe messages to the live communications servers and live OLTP system according to said determination.

2. A method for maintaining a high availability OLTP system having a live and shadow OLTP system including live and shadow WANs, live and shadow communications servers and remote client computers, said method comprising the steps of:

a) establishing a connection between the live and shadow OLTP systems;

b) establishing a direct connection between the client computers and their respective live and shadow communications servers;

c) appending client computer identification data to the input messages the client computers send to the live and shadow OLTP systems;

d) establishing a backbone network between the live and shadow OLTP systems by providing two dedicated communication circuits between the shadow and live OLTP systems;

e) maintaining a keep-alive timer in the shadow system;

f) transmitting keep-alive messages, via the backbone network, periodically from the live OLTP system to the shadow OLTP system, said periodic keep-alive messages being sent before the keep-alive timer in the shadow system expires; and

g) sending output messages from the live OLTP system to client computers via the backbone network, shadow OLTP system and shadow communications server to bypass potential live WAN and live communication server faults comprising the steps of;

i) having the live OLTP network manager accept output messages from the OLTP application program;

ii) having the live OLTP network manager create duplicate output messages using the client computer identification data contained in the connection list maintained by the live OLTP system;

iii) having the live OLTP network manager send output messages to the shadow OLTP system via the backbone network;

iv) having the shadow OLTP system relay the output messages to the shadow communications server which relays them to the client computer;

v) having the client computer network manager detect and discard duplicate output messages created by the live OLTP system; and

vi) having the client computer network manager pass one copy of the output message to the client computer application program.