Utilising high availability techniques

Phil Reakes explains why high availability fits better into a business continuity management strategy than DR does.

Everyone nowadays has heard of business continuity plans and many businesses have one in place. But if you don't have one you are in good company; Gartner research shows that fewer than 30 percent of the largest companies have invested in a full BCP. Even for those that do have a comprehensive plan, many rely on traditional disaster recovery methods to return to business-as-usual. The downside with this approach is that by the time the BCP has been implemented it may be too late; the business has already been impacted, sometimes terminally by an incident.

It's almost impossible for any company to mitigate against every possible IT failure as necessary in a disaster recovery plan. There is however an easy solution that isn't prone to human error - 'high-availability'. High availability can ensure your IT system recovers from almost any failure, normally without any data loss.

Many IT directors believe that a disaster recovery plan covers any eventuality. Disaster recovery is important but it should be a 'post-script' to high availability. Anyone dedicating time and attention to a disaster recovery plan without thinking about high availability will, effectively, be wasting valuable resources.

Planning, developing and testing an IT disaster recovery plan is a major exercise even for medium sized businesses. It is people-intensive, time-consuming and costly, and yields no benefit unless there is a disaster. Worse, even small changes to systems and applications can invalidate the plan without anyone realising until it is too late.

Anyone that fixes their sights on disaster recovery alone is going to take actions that will damage the company - because disaster recovery contradicts business continuity in some ways. To infer that business continuity equals disaster recovery is wrong. The best business continuity plan is one that avoids any situation of losing data. This can be achieved through a high availability infrastructure but not solely through a disaster recovery plan.

Investing in resources to implement a high availability infrastructure - protecting all operational data and applications from loss or failure - should be the first step in any company's business continuity planning. This will directly improve continuity by stopping localised hardware, software or communications problems from affecting the business. And beyond that, when a genuine catastrophe strikes, the disaster recovery plan for a high availability infrastructure is simpler to define, test, implement and maintain.

How do companies create high availability in their IT systems?
Implementing high availability revolves around managing the consequences of points of failure in the IT infrastructure. This is achieved by ensuring there is redundancy in the processing environment and adequate duplication of storage systems, as well as implementing very sophisticated fail-over for systems and data.

It is also essential to align the computing infrastructure with the corporate business strategy, while balancing return on investment against acceptable cost of risk. High availability recognises the stated business rules and simply adjusts priority levels of various applications. Ideally, high availability incorporates concept, hardware, software and bespoke automation scripting to provide the best resilience for the specific needs of the company.

A high availability management system knows not only what applications and resources are in operation, but also what the system relationships and business priorities are. It is constantly monitoring the vital signs of the system as a whole, watching out for possible problems. If a problem occurs - it could be a localised hardware failure, a system crash, even the loss of a whole server room - it quickly assesses what resources are left, and how best to restore any lost services. Sometimes it will recognise that there may be insufficient capacity in the live network for full restoration, and so it will use the defined business priorities and technical relationships to select which applications and service levels to restore. All of this will typically take a matter of minutes at the very most - and of course the people responsible for the system will be alerted with the precise issue so they can also take appropriate action.

The benefits of high availability
The goal of mission critical computing is guaranteed uptime to the customer (internal and external). The best high availability tools can provide a predictive, self-healing IT world, as a component can be fixed or fail-over procedures can begin before an outage or systems failure occurs.

Putting in place high availability can increase productivity, control and revenues as well as save time and resources through proactive monitoring and diagnostics. If true zero downtime is essential then fault tolerance is the requirement. But it is, and will remain, out of reach for the vast majority of businesses with even global corporations deploying it on a very small scale, if at all. By contrast, high availability with its combination of low entry and maintenance costs, virtual elimination of data loss, and prevention of high cost/high risk outages, is affordable for any company deploying corporate-grade servers and applications. As a result, many different organisations are now turning to high availability systems to protect their brand and bottom-line.

By adopting a high availability strategy and IT infrastructure, a company can prevent systems failure and downtime from occurring. Not only will revenues be stronger but it won't have to revert to its IT disaster recovery plan.

Even if a building is destroyed and other areas of a disaster recovery plan need to be implemented, an IT system with high availability across different locations will continue to operate without any catastrophic loss of data, enabling the business to survive.

Phil Reakes is managing director of Selway Moore.

•Date: 21st November 2003 •Region: Worldwide •Type: Article •Topic: IT continuity
Rate this article or make a comment - click here