Let the uptimes roll: why you need 24 x 7 availability
By Alan Arnold
Business requirements are not what they used to be. Driven by the increasing dependence on IT and the non-negotiable requirement for continuous online access, the definition of high availability (HA) is rapidly tightening. And what was once considered good enough availability is now unacceptable for most Global 2000 users, according to research firm META Group.
Indeed, it predicts that over the next three to five years, the holy grail of availability (true fault tolerance, defined as 100 percent continuous availability) will no longer be optional for most large, mission-critical, e-driven systems and not just those of the Global 2000, but also smaller and medium-sized enterprises (SMEs).
The new market forces shaping today s digital economy are hard task masters indeed: information is paramount, and constant, unwavering accessibility of that information is even more important. All this must happen in the totally transparent glass house of the e-business environment. Everything is revealed to customers via the unforgiving medium of the Internet and other mission-critical applications that are run throughout the enterprise.
Now, if your e-business site is down, even for a few minutes, you lose not only thousands of hits, but customers too if your customer doesn t get what he or she wants in three seconds, your competitors more reliable and more robust e-commerce site is only a click away.
The e-business phenomenon has its own set of rules, in which time to market is the be-all and end-all. Every business plan is perpetually on the verge of obsolescence. And the world runs on Internet time. This means that today there is only one time zone, and it s not EST or GMT, but NOW, believes Barry Hoard, IT writer.
It's simple to tell Internet time, he says. There s uptime and there s downtime. During uptime, companies transact business, serve their customers and plan for the future.
During downtime, he continues, companies watch their customers click their way to a competitor's Web site. Future? What future?
I couldn't have said it more succinctly myself.
E-commerce may be considered by cynics as merely the latest of the so-called New Age business drivers. Yet, when one examines the rules of this new game more closely, one sees that it is not so much the rules, as adherence to those rules, which seems to be setting the successes apart from the failures. Think only of the E-Bays, Amazons and Yahoos of the new world.
As these famous e-examples will testify, one of the most important of these essential e-economy rules is downtime of any form or flavor is an absolute no-no.
To put it bluntly, downtime is very damaging, not only to the bottom-line of a business, but also to the goodwill and public image of your enterprise. The latter is typically much harder to repair than any short-term economic losses you might suffer.
Scalability is a particularly acute problem, because Web sites must be able to accommodate spikes that come after promotional campaigns that spur increased interest in their products or services. Look at Computer.com.
Two days after the start-up spent $3 million on advertising during Super Bowl 2000, half a million people flocked to its Web site which did not crash.
All this fanatical dedication to uptime is well-founded. With so many Web-based opportunities only a click away, customers won t wait for service. You've simply got to be up all the time. When a prospective customer is doing something on the Web, and the site is down, they will find another search engine or another retailer to buy from.
It's not enough just to have a reliable server, though. Hardware goes bad no matter how much high availability they build into it. This is particularly true for e-business, where you need a good infrastructure of many boxes, not just one. If a site goes down, you need to fail-over quickly.
As an example of how the rules of availability have changed, look at Charles Schwab. One of its maxims is to be worthy of its customers trust at all times. That means it has to be available whenever its customers want to do business.
This kind of situation makes you long for the days of your nice little data center with one mainframe. Now, the e-environment is very transparent and with the Internet, Charles Schwab s customers know when they re down, and how long they've been down.
The situation is particularly acute at Schwab, because more than just a number of its customers are in the media business.
Indeed, in sorting the prevailing e-business buzzwords and trends, high availability must, by its very nature, stand out as a top business imperative. How a company deals with its short- and long-term e-business strategies can have dire consequences if done incorrectly. Conversely, a solid, well-thought out e-business strategy will help a company stand out from its competition and contribute as a major differentiator. Think about this for a moment and you will realize that, of the millions of Web sites in the world, you can probably only think of a few sites that you visit regularly. Why do you visit them? Probably because they are convenient, serve your needs, and are available when you need them.
In the years to come, veterans of the e-business economy will look back at a business landscape where high availability was one of the most outstanding features and enduring business practices of its time, and they will wonder why it took so long for people to realize its significance. Any world-class company with a business on the Web (or off the Web, for that matter) will have a world-class high availability scheme to ensure its systems are always running when planned.
Why? What is the business case for this seemingly essential practice, and why, as a modern business person participating in this much-vaunted new economy, should you even be giving it financial and practical consideration in your business environment?
First off, decide where your company fits in. Do you want to remain confined by the old economy paradigms, as restrictive as the bricks and mortar which housed them? Or are you prepared to take a deep breath and make that quantum leap into the new economy?
If your answer is yes to the latter, then you must assume all the responsibilities of being a global player in the new global economy.
Typically, businesses of today are not only global in the virtual sense of the word, but also in terms of their physical and geographic presence. Location simply doesn't matter in the new global economy of the e-business model. It therefore makes good business sense to examine very carefully how to leverage your existing infrastructures and systems to deliver maximum return on investment in the most effective and efficient way possible.
In order to do this, it becomes necessary to employ one or more methods of ensuring constant high availability, or maximum availability of those systems and resources, across the enterprise.
High availability is defined as a high degree of uninterrupted access to an organization s computing resources and data files. This involves the management of planned and unplanned downtime and making information available in the format users requires it be they employers, trading partners or customers.
Ideally, best practices in high availability will ensure 24 x 7 x 365 operation of computer systems, even when information is shared across multiple computing platforms. This means an organization must understand where it is vulnerable from a systems perspective and plan accordingly. All systems have different methods that can help to ensure their environment remains highly available to their end user community.
Simply put, you must ensure your computer hardware, software, applications and infrastructure will support your company in case of a failure. This is where you have to do some careful, business-orientated thinking. Ah, return-on-investment, I hear you say. Yes, but, that s not the whole story.
While higher levels of availability increase the investments required, it is a good idea to have a good understanding of the financial value IT systems provide to the business, as well as the costs to the business if these systems are not available.
Analysis in this area is often delayed as it is time-consuming and difficult, considering the number of variables that exists within the organization. However, once it has been determined, you have an invaluable tool that can be used to establish availability requirements, help you cost-justify appropriate investments in availability management solutions and assist you to measure all returns on that investment.
Despite the many advances toward high availability, many companies have no business continuity plans for their Internet-based applications. Taking a look at a recent version of the Vulnerability Index, a bi-annual study of large computer users, on a scale of 1-100, with 100 representing maximum vulnerability, the average Internet Vulnerability Index score for organizations participating in the study was an alarming 66.
In comparison, the average score for LANs was 55, while enterprise and application servers registered an index reading of 42. Overall, the study found that only 39 percent of companies had an effective business continuity program in place for all three of these mission-critical systems.
The cost of downtime varies from company to company. Contingency Planning and Research estimates it at $6.5 million an hour at retail brokerage houses. For brokerages, downtime is an unmitigated disaster, but even for dot.coms, once you factor in the legal costs of lawsuits and the impact of reduced stock value, the downtime estimates only go up. Furthermore, the impact of downtime often exceeds hard cash, going straight to corporate image. It's about how quickly you can lose a customer s trust and it s hard to put a price tag on that.
The objective of planning is to provide businesses with the knowledge and tools that will help them develop or refine an availability plan that will balance the cost of a failure with the cost of protection against that failure.
It will help answer questions, such as how to decide which recovery facilities are worth implementing now, which should be considered later, and which don t apply at all.
So, with all these dire warnings, why is it that so many companies are still not deploying protection against catastrophic downtime?
One reason could be CIOs and other IT leaders are overwhelmed by their daily tasks of keeping production systems up and running. Consider some of their challenges: the time-to-market demands of e-business are compressing the cycles for application projects, while backlogs are piling higher. Good technical people are harder than ever to find. Top management is focused on revenues and expects 100 percent business continuity.
CIOs are so knee-deep in other projects that they simply rationalize the chance of having a disaster is negligible. That may be so, but publicly traded e-commerce companies who fail to protect themselves with business continuity and high availability are taking a myopic approach to a potentially fatal problem, warns Peter S. Cohan Associates, a consulting and venture capital firm.
It's not high enough on their list of priorities, and top management is not taking the issues seriously, he says. Many sites are just trying to keep up with their competition, and it will take a disaster to wake them up.
When high availability demands 24 x 7 uptime on a global basis, there is no good time for shutting systems down. In April 1999, a software upgrade cost AT&T $40 million in rebates. The way out of this dilemma is straightforward: develop a solution based on a good business continuity plan. It may not seem like a lot to ask, but, as the Vulnerability Index reveals, at many major companies, it simply isn t happening.
Companies unwilling to spend money on continuity services might view the situation in a different light if they considered the risk to their revenues. These companies should ask themselves if it makes sense to spend one percent of their revenues to ensure the safety of the other 99 percent. When the question is phrased that way, it s hard to come up with no as an answer.
The real cost of downtime, opportunity cost, is perhaps the hardest form of loss to quantify or calculate but it s also the hardest to recover from on a long-term basis. Bad news travels fast and, in the relentless transparency of the new economy, modern businesses would do well to remember that it travels even faster.
Recapturing customer loyalty is far more difficult than attracting the new customer in the first place. Your company must do everything possible to ensure you don t lose a customer because of an unavailable system or application. This is simply not an option for any organization that considers itself world-class.
Alan Arnold is EVP of Engineering and Research, Vision Solutions Inc.
Copyright 2001, availability.com. Reprinted by permission.
FOR MORE INFORMATION: