At the end of last term, the College announced a $22 million donation from an alumnus to establish a new institute for computational science. It is sadly ironic that at the same time, the school began to experience several weeks of network problems. Internet outages, BlitzMail crashes, power failures and GreenPrint problems abounded. It seemed every part of our network was starting to fall apart at once. As the administration was trumpeting Dartmouth's "legacy of leadership in computing," students were panicking as web resources became inaccessible and paper deadlines drew near. Indeed, some students probably felt themselves cut off from civilization when thefacebook.com became inaccessible.
Dartmouth does not deserve such a shoddy network. The entire campus is connected to the internet by one router. When that single machine fails, our connection to the internet is lost. Perhaps even more condemning is that fact that the uninterrupted power supply (UPS) units themselves failed when the main power went out. These UPS units are supposed to act as batteries to keep the servers going while the main power supply is being restored. That the emergency backup power supply failed is an indication of the unreliability of our network.
Our present network remains vulnerable. Right now, Dartmouth is connected to the internet by one router and one internet provider. If either one fails, our access to the internet is lost. The digital library becomes nothing more than a pretty website with a lot of dead links. The reliability and stability of our network must be improved. It is unacceptable that one machine failure can shut down the hole system.
We need a redundant network with the excess capacity and capability to manage problems without causing the failure of the entire network itself. It does not matter if one router fails as long as there is a second one available to take up the slack. This means having at least two routers and several connections to the internet so that any one part of the network can fail but the entire network will remain online. This means investing in new equipment and bandwidth to make our network more reliable and stable.
Redundancy is not a waste of resources. The web-hosting industry, which provides internet support for large corporate clients, is built on redundancy. Web-hosting companies pride themselves on the redundancy of their networks. The basic premise is simple: the network cannot fail. When a network fails, this means lost business, lost e-mails and lost prestige for client companies. To avoid this, web-hosting companies invest in redundancy, buying enough equipment to operate them at half capacity, just to prepare for the eventuality of a machine failure. Diesel generators are not uncommon to serve as back-ups to UPS units.
The Computer Age has brought great benefits, but computers are machines and machines can fail. Network problems are not rare events and therefore should be anticipated. It is for this reason that students are always warned to backup their papers, "just in case." It seems that Kiewit should listen to its own advice. At the last network outage, The Dartmouth (March 5, "Internet Outage Causes Headaches for Students") reported that Kiewit had to wait for a new router to be shipped to Hanover before the problem could be resolved. In fact, the article suggested that Kiewit was aware that the old router was problematic but was in no hurry to replace it until summer.
Such complacency is unacceptable. Kiewit is apparently unaware that their failures have a real consequence on the lives of students. After the problems with the BlitzMail servers, Kiewit sent out a message to all students apologizing for the "inconvenience" caused. Having to walk up five flights of stairs because the elevator is under repairs qualifies as an "inconvenience," but losing your e-mails and having your e-mail account turned upside down like a frat basement on Homecoming weekend is no longer an "inconvenience" but a real disaster. Even then, it seems that students are expected to be "patient and understanding" because no one in Kiewit checked to make sure that those emergency batteries actually worked as they should have.
Kiewit should realize that they are providing a critical service to this College. Kiewit has to take responsibility for the series of network problems and work to preempt them in future. Network problems should not be accepted with an air of resignation and dismissed as mild "inconveniences." These are real problems and have a real impact on the lives of students who are the real consumers here. And I must say customer service has been very bad these last few weeks.
