Blueprints for High Availability: Amazon.co.uk: Evan Marcus, Hal Stern: Books

Frequently Bought Together

More About the Authors

Discover books, learn about writers, and more.

Evan Marcus

Hal Stern

Product Description

Expert techniques for designing your system to achieve maximum availability and predictable downtime

With your company′s reputation and profits at stake, downtime on your 24/7 web site is not an option, nor is poor application performance. Now in its second edition, this authoritative book provides you with the design blueprints to maximize your system availability.

Striking a balance between costs and benefits, the authors show you all of the elements of your computer system that can fail–as well as ways to assess their reliability and attain resiliency and high availability for each one. A unique feature is "Tales from the Field," a collection of true–to–life experiences that will help you avoid mistakes and deploy your system with confidence.

Learn how to design your system to limit the impact of such problems as computer viruses, natural disasters, or the corruption of critical files and discover how to:
∗ Implement effective backup–and–restore and tape management strategies
∗ Arrange disks and disk arrays to avoid downtime caused by inevitable failures
∗ Utilize technologies such as Storage Area Networks (SANs), Network Attached Storage (NAS), Virtualization, and clustering
∗ Achieve effective application recovery after any part of the system has failed
∗ Replicate critical data to remote systems across a network

From the Author

High Availability is an end to end proposition. Our book starts from a very simple premise; we call it our mission statement: "You cannot achieve High Availability (HA) by simply installing failover software, and walking away." Highly available systems are predictable, that is, they behave with consistent response time and recovery time, and can be managed consistently, as the load placed on them increases. This level of predictable, continuous computing can only be achieved by solving the problem at multiple layers.

In this book, we start from the core of a single system and work our way "out" to an end to end approach. At the heart of the design is failover software and disk redundancy techniques. Failover software, or HA software, such as Microsoft Cluster Server, VERITAS Cluster Server, or Sun Cluster, quickly and smoothly transfers control from a failed computer to a pre designated standby, allowing it to take over the functions of the first. From there, we cover network design, network services such as naming, directory and file systems, and system software functions like databases and web servers that run on these highly available platforms. We also talk about replication and application-level issues that require careful consideration.

Failovers will not help you recover from poor system or network performance, network outages, buggy software, full disks and file systems, human error, poor design, and a bunch of other potential pitfalls. We take a detailed look these and other causes of system downtime. We discuss protective measures that you can take to eliminate them, or at least to minimize their impact. Since some system outages are inevitable, we outline design rules for your systems so that you know how long each common kind outage will last, and so you can set user expectations accordingly. Our goal is to let you look at the problem of availability from the perspective of a developer, a systems manager, an architect, or, yes, even a technical manager.

The book is neither a Windows or a Unix book. Our advice, and most of our examples, apply equally to both environments. Good design techniques apply to all environments.

Thank you for your interest in our book. We are very proud of it, and we believe that it can help you design systems that deliver on the promise of increased availability. --This text refers to an out of print or unavailable edition of this title.

See all Product Description

Inside This Book (Learn More)

First Sentence
Despite predictions in the 1970s (and extending into the 1980s) that computers would make everyone's lives easier and give us all more leisure time, just the opposite seems to be taking place. Read the first page

Explore More
Concordance