In the 10 years that he’s worked in senior technology roles at SBLI USA, Paul Capizzi says the company’s only outage of its customer-facing websites occurred in August 2003, when the U.S. Northeast suffered a widespread blackout.
“We had a good plan in place, but when the power is down for that long and you don’t have a redundant data center, there’s only so much you can do,” says Capizzi, vice president of technology for the New York City-based life insurance company. Fully preventing outages isn’t possible, yet with proper planning and careful selection of technology and hosting partners, a small business can protect itself against unnecessary downtime and be able to quickly restore service.
As an SMB competing against much larger life insurance providers in New York, SBLI must provide the same measures of reliability and customer data protection as the big guys in its market. The company hosts its own sites and data center and has two ISPs, backup Web servers and battery backup power. “We always have had a very redundant environment,” Capizzi says. “We feel that this is critical for the demand on our sites and to support our 300,000 and growing customers.”
Whether your company requires this much infrastructure to prevent an outage depends on how critical online operations are to revenues. Do you know what an hour or more of downtime could cost your business in terms of lost sales or customer defections? If not, do the analysis. At a minimum, your company can request a “static page” service from your Web hosting company, Capizzi advises. This service displays a customized, branded message to visitors who reach your page if the site is down, instead of the generic error page.
Preparing for and dealing with Web outages
If your site is primarily used for lead generation, it’s probably not a disaster if it goes down for a few hours or even a day. But if you’re conducting business transactions from your site, a couple hours of downtime could cost thousands of dollars in lost revenue or customers, not including the time and money spent to restore the site. Here are some prevention strategies that work:
1. Backups. Performing a regular (daily) backup of all systems that your website accesses is a basic requirement. But part of your site’s data will reside externally on the servers of the hosting provider. “They’ll all say they back up the site, but you need to test them,” advises Peter Connolly, owner of Web design and consulting company KP Direction.
To do so, upload a file to your site, then delete it a few minutes later. Now ask the hosting company to retrieve it for you. This should happen in about 30 minutes or less, says Connolly. If not, that’s a red flag. Regardless, experts advise backing up your entire website on your own, despite what the hosting company is doing. You can use free software such as FileZilla to copy all of the data and files to your own internal systems through the website control panel, says Steve Glass, a San Diego-based Web developer.
2. Hosting options. There’s no such thing as 100 percent or even near 100 percent uptime, so choose a provider you can trust. If a provider isn’t easy to work with, dump the firm. “It’s better to have a responsive hosting company who listens to you and is prepared to work to get the level of service you need and can afford,” Connolly says.
Next, consider what level of service your business requires. With shared hosting, your site resides on a computer with many — even thousands — of other sites. While super affordable, if your site stores customer data or processes orders, you can’t use shared hosting because the security requirements aren’t strong enough, according to Connolly. At that level, you’ll need a virtual private server or VPS, a dedicated private server that’s securely partitioned on one physical server using virtualization technology. This will cost around $25 or $30 per month to start, he says.
At the far end of the hosting spectrum would be a dedicated physical server, which is necessary for high-volume sites receiving 1 million hits or more daily. Leasing a dedicated server may cost you $300 or more monthly.
The options don’t end there: Do you need dual hosting, meaning contracting with two ISPs or hosting companies? “If you truly need near 100 percent availability, you cannot — no matter what anyone tells you — single-source the hosting of your website on a single provider,” says Michael Dinn, former operations manager at telecommunications and hosting provider Internetworking Atlantic. “Going with a multi-vendor solution isn’t a simple undertaking, but gives you much better resilience against single points of failure.”
3. Logging and monitoring. Again, leaving everything up to Web hosting companies isn’t a good idea for revenue-generating sites. As mentioned, you can retain some level of control by doing your own backups. Maintain a manual log of what changes are made to the site, who made them, and when, all of which speeds the restoration of your site in case of an outage, Glass says. Also, use a website monitoring service, which emails you if your site is down at a particular location, advises Connolly. He recommends Hyperspin and Basic State, though the choices are vast.
4. Handling an outage. If your site suffers an outage, don’t panic. First, call your hosting company to find out what’s going on, and check sites such as this one to make sure the issue isn’t just your own. As soon as possible, notify customers about the problem and when you expect the site to be back in service. For a long outage, provide periodic updates to customers about status. Since restoring your site quickly and thoroughly can be a fairly technical process, if you don’t have technical expertise on staff, get to know a reputable IT consultant you can call in an emergency.