Breaking the Availability Barrier II
Achieving Century Uptimes with Active/Active Systems
by
Book Cover & Preview Text
Chapter 3: An Active/Active Primer
In the previous two chapters, we analyzed a concept that has the potential of revolutionizing computing architectures by providing extraordinary reliabilities at ordinary costs.
There are some classic parallels to systems with this extraordinary capability. When you pick up the telephone, you expect a dial tone. When you turn on a light, you expect electrical power.
Occasionally, a large group of users may be taken out of service as evidenced by telephone overloads during crisis times or by the two Northeastern United States blackouts in the last four decades. However, most outages are due to failures in the “final mile” – typically a tree limb falling on wires.
This extreme level of availability is achieved by having highly redundant networks that can route around failures. Active/active systems achieve their reliability by providing an equivalent flexibility; but they do so, of course, on a much smaller scale.
In the transaction-processing world, the classic approach to recovery from a processing system failure has been tape backup. However, recovery can be a very slow process and may take hours or days as the database is reconstructed. Besides, handling volumes of tape is an expensive and error-prone process. This has led to the use of active/standby system pairs in which the standby database is kept up-to-date via data replication. Such a configuration can reduce recovery time to minutes or hours but requires purchasing a second system whose capacity is not available for the application.
Active/active systems represent a major paradigm shift for high availability systems. No idle capacity need be purchased for backup purposes, and recovery from a failure can be completed in seconds.
If active/active systems represent such a powerful approach to system architecture, why are they not evident in all critical systems today? The answer is that there are, in fact, many active/active installations today experiencing almost no downtime. However, these are generally one-of-a-kind implementations that have been custom designed and built. Until recently, there have been no generic solutions to building active/active systems. However, generic solutions are now beginning to appear; and with these will come an increasing number of such systems.
Formats
Book Details
About the Book
This book is Volume 2 of a three-part series on active/active systems. It describes techniques that can be used today for extending system failure times from years to centuries, often at little or no additional cost.
As our daily lives and corporate well-being become more dependent upon computers, system reliability grows increasingly important. No longer are frequent system outages acceptable. In many cases, failure intervals must now be measured in centuries.
Starting with a summary of Volume 1, techniques for achieving extraordinary availabilities are reviewed. These techniques use active/active architectures, in which multiple independent nodes using a common distributed database are cooperating in a common application. Should a node fail, all that is required is to switch the users on that node to a surviving node.
Equally important to the achievement of high availability is the ability to upgrade the system hardware and software without denying service to the users. The procedures to do this within an active/active system are described.
The secret to high availability is to let it fail, but fix it fast. This volume explores the server, database, and network redundancy techniques that allow fast-fix to happen. The cost considerations involved in such redundant architectures are also explored.
About the Author
Dr. Bill Highleyman, Paul J. Holenstein, and Dr. Bruce Holenstein have a combined experience of over 90 years in the implementation of fault-tolerant, highly available computing systems. This experience ranges from the early days of custom redundant systems to today’s fault-tolerant offerings from HP (NonStop) and Stratus.
Dr. Bill Highleyman has done extensive work on the effect of failure mode reduction on system availability. He has built fault-tolerant systems for train control, racetrack wagering, securities trading, message communication, and other applications. He is the Managing Editor of the Availability Digest (availabilitydigest.com).
Paul J. Holenstein and Dr. Bruce Holenstein have architected and implemented the various data replication techniques required for the availability enhancements described in this book. Their company, Gravic, provides the Shadowbase line of data replication products to the fault-tolerant community.