- The 10 dumbest mistakes network managers make
- Six Windows 7 features admins will actually care about
- Why the iPhone can't be "killed"
- Nortel enterprise chief wants to bring back Bay
- More porn sneaks onto the iPhone
Four years ago, when the New England School of Law in Boston suffered a total network failure that lasted almost a week, the IT department knew a major overhaul was in order.
From two cramped server rooms, the school's small IT staff was trying its best to maintain a wireless network for 1,100 students, e-mail, VoIP phones, payroll applications, Web servers, FTP servers, an ERP system, and specialty applications that handle building security functions. But there were too many single points of failure, and when the school's main router and Web server failed at the same time, the results were disastrous.
"Even though we didn't have too many pieces fail, they were expensive pieces and they were pieces the rest of the network couldn't run without," says Nathan Reisdorff, director of IT at the school, which is changing its name to New England Law/Boston.
The school's dean, "staring at 1,100 angry students" in Reisdorff's words, was quick to propose an overhaul of the college's network. Besides suffering from single points of failure, New England Law was running out of space in the two server rooms and having trouble getting enough air conditioning to the servers.
The keys to overhauling the network, making it more reliable and cost-efficient, were VMware's virtualization tools including VMotion along with HP blade servers, storage-area network and storage virtualization products. Over the past year and a half Reisdorff and CIO Charles Killam built a redundant network with automatic failover and backup capabilities, while shipping about 20 legacy servers out the door and reducing demand on power and cooling by about 30%. (Compare server products.) But the school experienced some fits and starts and new problems before finding the right strategy.
Reisdorff explains that he and his colleagues tried to implement a best-of-breed approach with an EMC storage-area network (SAN) and Dell servers, but never got everything fully implemented because of equipment problems and a lack of experience designing a high availability network. Servers failed, there were problems hooking servers up to the EMC Clariion CX300, and when New England Law went back to the providers, they simply pointed fingers at each other, according to Reisdorff. Unforeseen costs plagued the project almost from the start.
Partner Content
Explore the Ultrium Edge
The powerful tape technology can address data security with tape encryption as well as long term data protection.
Find Out More
Disk and Tape Square Off
Discover what disk and tape really cost and which solution provides lower total cost of ownership and optimizes energy use for your organization
Download this White Paper
Don't Fall for the Myths
The Clipper Group explores the truth behind the myths of tape, digging into the misconceptions in the disk vs. tape debate.
Review this information
information examination
An examination of information security issues, methods and securing data with LTO-4 tape drive encryption
Read this analysis
Comments (3)
My thoughts as wellBy Anonymous on January 7, 2009, 8:12 amI thought the same thing, where was the actual NETWORK failure information? True, there was a passing mention at the beginning of the article: "and when the school's...
Reply | Read entire comment
Bad Title to ArticleBy Jon S. on January 6, 2009, 10:35 pmSo after reading this, my understanding is that other than a vague comment about 1100 wireless users, there were actually no problems with their NETWORK, but with...
Reply | Read entire comment
No downtime for EVA drive failuresBy Anonymous on January 6, 2009, 4:48 pm"The EVA has all the RAID built in, so if there's a bad drive it will move the information off to different drives so you can replace it without any downtime," That's...
Reply | Read entire comment
View all comments