By Jim Wagner
Executives at Microsoft Corp., a company famous for its
employees pulling all-nighters, are likely going to be doing the
same this weekend as they try to figure out what went wrong.
The software giant had what any high-tech organization would
deem “a bad week,” with repeated Domain Name Servers problems that
caused system-wide shutdowns throughout the week, at one time for
as long as 23 hours. As a result, Microsoft Web sites worldwide
went offline.
It’s a problem that might lead the company to scrap its $200
million ad campaign promoting .Net services, announced Monday, and
work to revamp its entire network infrastructure.
.Net service is a combined package of its Microsoft Office,
Visual Studio and bCentral, and the foundation for what Microsoft
planned as an empire moving from the PC to the Internet.
In the future, officials had planned on applications running
over the Web, instead of through the home computer’s hard drive,
thus ensuring the company’s existence well into the first decade of
the 21st century.
Rob Enderle, an analyst at the Giga Information Group, said this
week’s setbacks effectively destroyed its still-fresh campaign and
any Web dynasty hopes.
“It’s destroyed,” Enderle said. “While they’re running a big
campaign talking about the reliability of MS products is not the
time to have major outages at the site, particularly for sites as
visible as Microsoft. (The outages) pretty much destroyed their
advertising campaign, and any value they might have achieved from
that campaign is pretty much gone.
“In fact, there’s even a risk that the campaign will become an
industry joke,” Enderle said
Microsoft officials still claim a technician, who incorrectly
configured its network of DNS servers, was to blame for the
blackout Tuesday night and Wednesday of popular sites like MSN.com,
MSNBC.com, Encarta.com and Hotmail.com. Even after the DNS servers
were fixed, many customers still reported spotty reliability
getting onto a Microsoft Web site.
That, in turn, officials say, weakened its network enough for
malicious hackers, called crackers, to find to blitz the network
Thursday with a Denial of Service attack that brought the site to
its knees for the second time this week. Microsoft reports its
network was down about five hours before the sites were
restored.
Microsoft placed a call to Federal Bureau of Investigations
after the DoS, although its uncertain what can be done.
According to a member of the FBI’s press office, there is not
much that can be done after the attack is over with, although
agents take the report and run a preliminary investigation.
Industry and security analysts were left shaking their heads
after learning Microsoft kept all its DNS servers running on one IP
subnet on the same network. The setup ensured that a DoS attack, or
common system failure, could bring down Microsoft’s entire
collection of Web sites.
“This showed an exposure that should not have existed at
Microsoft,” Enderle said. “It indicates that there is a critical
problem that needs to be repaired in the way they laid out their
entire network and it will take them a while to come up with a plan
to not only address this exposure but to address other exposures
that are likely to occur.
“What basically happened,” Enderle continued, “is that somebody
was sleeping on the switch and probably was for some time and the
end result is they undoubtedly have to change a large part of their
infrastructure. And the first part of doing that is to come up with
a plan so that they’re not creating more problems in the
process.”
Microsoft officials were unavailable for comment.