Do you have the right e-commerce applications?

By Neil Plotnick 30 Aug 2001 | availability.com

Digg This!     StumbleUpon     Del.icio.us

Consistent availability is the single most important objective for any company relying on Internet technologies for e-commerce. Every segment of the system's architecture, from operating systems and applications to hardware and communication links, must be designed with resiliency and fault tolerance in mind. Unreachable Web sites or crippled back-end database servers can immobilize operations and send customers off in search of other vendors. With rivals often only a mouse click away, steps must be taken to ensure that your company's systems are always reachable.

A secondary goal must be the ability of your systems to respond to varying data loads. Supporting extremely active customers and complex transactions demands faster servers and pipelines. Customers or business partners that are denied access due to system saturation will likely seek alternatives -- especially if performance is spotty. Excess capacity should be provided to maintain a positive customer experience, even under peak usage conditions.

Thankfully, there are many existing mature technologies that promote maximum system availability. Fault-tolerant hardware such as disk arrays and power protection provided by an uninterruptible power supply (UPS) are two common schemes. Clustering and automatic fail-over are increasingly becoming standard features of modern operating systems. The capability for multiple computers to work as a team and automatically assume the roles of a failed system greatly improves overall reliability. By distributing loads among multiple computers, capacity and performance is also boosted.

Paying the price

As greater availability is added to your infrastructure, costs will rise. You must determine the level of reliability your company demands when designing a solution. Many e-commerce firms mirror their data and applications at multiple locations. While multiple, redundant data centers can help reach the elusive goal of 100 percent uptime, the cost of building and maintaining duplicate sites can be daunting. Constant monitoring of system performance is necessary to respond to any problems that arise. Thus, staffing and equipment to provide 24/7 coverage is also required. Finally, vendors usually demand a premium for enhanced service and support contracts that provide the quickest response times.

While operating systems, such as Windows 2000, and hardware are available with fault-tolerant features, many applications are not similarly equipped. Increasingly, fault-tolerant capabilities are planned for the next generation of e-commerce applications. However, some vendors do have existing support for these features; for example, Sun-Netscape Alliance's iPlanet Application Server has always been designed for high-availability architectures.

Seeking outside help

Corporations must often decide between designing and maintaining their own solutions or relying on others to build and host their data and applications. The decision to use either business model is based on a variety of factors. Basically, you must decide where to place the burden for maximizing uptime.

Expertise and experience are two excellent reasons to partner with an outside vendor. Consider that the types of applications and systems required for e-commerce are among the most complex available and consequently more difficult to manage. It can be costly and time consuming to attempt to train your existing staff to master new technologies and systems. Specific programming or other talent may be difficult to hire, especially given the incredibly high demand for Internet-savvy workers. Often the fastest way to acquire necessary abilities is to simply "lease" experts from outside firms. Relying on outside experts will also allow your company to take advantage of both the positive and negative experiences they have had.

There can be tremendous competitive advantages realized by using hosted applications and data. Your company can rid itself of the burden of maintaining complex and expensive servers. This is also a way to avoid the inevitable upgrade cycles that these systems require every few years due to obsolescence. Internet technologies continue to evolve rapidly, so a long-term investment in equipment is often a bad decision.

Reading the fine print

When relying on the services of an outside firm, pay careful attention to the service guarantees. Be certain that contractual obligations are clear and commitments to reliability are well understood by all parties. If possible, take a tour of the vendor facilities and verify that they are following the best practices for data back up and system redundancy. Verify that "hot spares" and other measures are provided in case disaster recovery is necessary.

Sometimes the desire to manage the entire project internally outweighs all other considerations. Companies with strong IT departments and extensive infrastructures may feel comfortable undertaking new challenges and systems. Prior negative experiences with consultants may reinforce this position. Only the most capable companies can assume responsibility for the availability of the complex applications and platforms required by powerful e-commerce products.

Whatever solution you ultimately rely on, never ignore the importance of testing disaster recovery procedures. Equip your IT workers with action plans that specify which steps to take when certain errors occur. Make sure that contact information for your various service providers is up-to-date and readily available. A calm, measured approach is only achieved when you have confidence in your system information and have practiced managing system snafus.

Neil Plotnick is the author of "The IT Professional's Guide to Managing Systems, Vendors and End Users" and a former columnist for PC Week.

Discuss availability and uptime with your peers

Pose a question to our app server guru Jesse Feiler

Tags: Availability & QoS,  VIEW ALL TAGS

Digg This!     StumbleUpon     Del.icio.us

RELATED CONTENT Availability & QoS Disaster recovery: know what you really need Security and disaster recovery Just how available should production data be? New Stratus server won't put up with downtime How to make the most of performance management tools Let the uptimes roll: why you need 24 x 7 availability Web capacity planning: A delicate balancing act A backup bind Can Windows take on the enterprise back end? Figuring out the metrics of "five-nines"

RELATED RESOURCES 2020software.com, trial software downloads for accounting software, ERP software, CRM software and business software systems Search Bitpipe.com for the latest white papers and business webcasts Whatis.com, the online computer dictionary