Managed availability in a cross-platform environment: mistakes to avoid when planning business continuity

Computer Technology Review,  Nov, 2004  by David Wegman

• E-mail
• Print
• Link

Financial administration and overseas CRM are on Linux; shop-floor operations and automated fabrication are on iSeries; and the front-end, browser-based eBusiness on Windows knits together your supply chain accelerating customer contact with real-time lead generation that's e-mail blasted to the sales territories.

Sounds impressive, but it all means nothing if it's not available to users when, where and how they need it.

The issue is information availability, a business requirement that's increasingly gained attention since an operator error crashed the first server and left operations treading water while management asked, "What happened? How long will it take to fix this? How much is this costing us?" Plenty then--and more now.

There is a lot of data available, but averages from a number of analyst firms reveal downtime costs between $250,000 to over $3.6 million per hour across several industries and, in some cases, as high as $15 million per hour. Some analysts have recently estimated that U.S. businesses lost over $9 billion dollars due to downtime in 2003 alone.

[FIGURE OMITTED]

Never before have information assets been so critical in leveraging competitive advantage, driving revenues and profitability, and sustaining business viability. This growing importance of information availability is the byproduct of a mega-change in the importance of data--its accumulation and how it is processed, accessed, shared and transformed for competitive advantage. Elements of this tide of change include:

* The evolution from highly centralized computing, in which a limited number of computers and individuals provided the information needs of an organization, to a highly diversified approach, whereby huge databases might be centrally managed but the applications and need for the data is spread throughout the nodes of a networked or clustered enterprise.

* An almost incomprehensible explosion of information reflected in the current worldwide production of between two and four exabytes of brand-new data annually. This equates to over 400 megabytes of brand new data every year for every man, woman and child on the planet.

* The conversion of business applications and activities, such as e-mail, from workplace conveniences to critical dependencies that anchor information sharing and integration.

* The emergence of the 24X7 business day in response to expanding worldwide commerce, diversified supply chains, remote employees with flexible work schedules, as well as expanding e-business practices.

* The increasing complexity and interdependency of both hardware and software architectures characterized by various best-of-breed and legacy applications interacting across mixed platforms and databases in an inter-dependent enterprise, which increases the risk in component and therefore system failure.

Business needs have changed and the users have driven an accelerating tide of change, which has altered our perspective on availability both in terms of technology and methodology. The common notion of "High Availability" has actually evolved into a broader business discipline called "Managed Availability" which is best defined as the ability of an organization to deliver consistent, predictable access to information for any user wherever, whenever and however they need it. It is not 24X7, but a business deliverable that has specific service levels expected by the organization and one that must be defined by flexibility, scalability and a "variable tolerance" for planned and unplanned downtime.

In today's world, High Availability is actually a subset of the broader Managed Availability, along with Disaster Recovery, Improved Availability and Continuous Availability. This may be charted across an Availability Continuum (see Figure) and these "degrees" of availability are defined by the particular emphasis each user gives to data and/or applications recovery. Disaster Recovery, for instance, is primarily a defensive or data-protective action offering limited but specific data recovery points, and therefore it is said to be "reactive to business requirements."

Being "proactive" moves us to High Availability where the operational needs of an organization (the applications) take center stage with an assumption that all the data is intact. The focus is on the time it takes to recover certain system components and assigns priority to applications availability over data availability.

As you can see, one size of a solution, especially in cross-platform scenarios, does not fit all; and it is clear that in between the two extremes of the top-right and bottom-left of the Availability Continuum there are many, many variations of solutions. Your Managed Availability decision should be influenced by what best fits the needs of the present business requirements coupled with the degree to which the business may evolve and how you will manage that change. And under it all is how the solution can be cost-justified.