Web Outages Raise Reliability, Security Questions

Extensive network outages at several of Microsoft’s Web properties between Jan. 23 and Jan. 26 have raised questions about the company’s policies and practices in the areas of security, Web architecture, change management, and crisis management. None of the outages were caused by flaws in its products, and similar problems have struck many other companies with large-scale Web operations. Nonetheless, the outages hurt Microsoft’s image just as the company is pushing itself as a maker of reliable enterprise products and the leading proponent of moving essential business services and applications to the Web.

DNS Failure Takes Sites Off Net

According to Microsoft, at about 6:30 p.m. on Jan. 22, a technician made an incorrect configuration change to the routers between the company’s Domain Name System (DNS) servers and the Internet. A DNS server translates a Web browser’s request for a simple-to-remember URL, like www.microsoft.com, into the numeric Internet Protocol (IP) address of a Web server where the site actually resides. The configuration change blocked requests to Microsoft’s DNS servers, meaning that users could not access certain Microsoft Web sites unless they happened to know their IP addresses. Affected properties included MSN (including subsidiaries like Hotmail and MSNBC); Microsoft.com; and other Microsoft-operated sites, such as Expedia and CarPoint. (For details on the DNS system and how the outage interfered with it, see the illustration “How the Outages Happened“.)