Preparing Your Site for Speed and Reliability
By Mitch Wyle
As Web sites become the foundation for critical interactions with customers, employees, partners, and suppliers, site performance and availability become measures of business viability, not just technical quality. In an environment where usage patterns are changing daily and new services are constantly emerging, the speed and reliability of network service are critical. However, daily press reports of spectacular outages and failures -- as well as our own daily frustrations surfing the Web -- indicate that the Internet as a whole has a long way to go to achieve the reliability of other business systems such as the telephone.
For the purposes of this two-part article, we'll consider two problems solved by using broadband Internet connections and implementing load balancing. In this installment we'll examine the distribution problem. Many of us need to make software, large files, or streaming media available to a large number of people over the Web. For these applications, we require very high bandwidth, but not necessarily a lot of computing power or low latency in each packet of the connection. In the upcoming article, we'll deal with the common (and somewhat sought-after) problem of receiving too many hits or transactions on your Web sites.
ISP and Data Center Selection
Before you can deal with either the traffic or the distribution problem, you must optimize for peering and verify redundant routes over more than two major Internet backbones.
