From Computer Business Review, a sister publication
It is more than a little ironic that the key plank of the network computer concept is also the major barrier to its deployment today: bandwidth. The NC is useless without a fast, low-cost, consistently available Internet/Intranet connection – the place where all its applications and data are stored. NC proponents argue that most corporations already have the necessary internal high-speed network connections in place to support the NC mode of working, which is why it is initially being pushed not as an Internet device, but as an Intranet device. We’ve been demo-ing it on ethernet. You can plug NCs into a PC LAN without having to do anything special, says Oracle CEO Larry Ellison. But there are still a number of concerns. Users will have to wait for a network connection before they can get any work done, say the PC suppliers. What will they do if the network or the server goes down, or if they are denied access to files? I’d argue that user files are much safer with me. I can guarantee the integrity of files. We [Oracle on the server] are much more secure than the PC, 100 times more, 1,000 times more, counters Ellison. Another major concern is control over data access. How will network administrators ensure that the right users get access to the right applications and data? Oracle says it will be equipping its devices with a smart card interface to identify and authorize individual NC users. It will include details of the name and password of the NC user together with information of authorized services, says Ellison. It is not clear if this will cost extra. The challenges associated with the home market, and business users out of the office, however, are much more taxing – and it is in this market that users are most likely to be attracted by the low price and the centralized security. While the majority of corporations already have high speed internal networks in place, most countries, with the exception of North America (which is largely cabled), still operate on slow, narrowband telephone networks which are incapable of supporting the data types, or speed of data transmission, required by network computers. It is a transitory problem, says McNealy. Don’t bet against bandwidth. The world is quickly moving to cable modems. Think of all the fiber that is currently being laid – gizillions of gigabit miles, he says. McNealy and Ellison’s hope is that competition to provide Internet access services between the telephone and cable companies, and the continued global deregulization of the telecommunications industry, will lead to the worldwide deployment of broadband, high-speed networks; and that current generation 28,800 bits per second modems will quickly be replaced with cable modems, capable of transmitting data at 4 million bits per second. This is a realistic vision, but it won’t help sales much this year or next. Until that happens, NCs will have to employ sophisticated data compression technology to attempt to speed up data access and delivery. There are also significant telecommunications pricing issues. In the US, local calls are all but free but line charges in Europe can make running applications and data on the Internet very expensive. This is a potential issue but the second you get real competition in Europe you’ll get lower costs on a per session basis, says Ellison. I believe that in the US you’ll soon have a flat fee irregardless of how many calls you make, where you call and how long you stay on the line. In the meantime, says Ellison, users will be able to minimize the amount of time they are actually on line through the use of flash cache. Undoubtedly these issues will be addressed – but probably a lot more slowly than NC proponents suggest. According to Blatnik of Zona Research, the bandwidth issues won’t be resolved anytime soon. Deregulation and competition in North America makes the business model feasible but, he says, in the rest of the world there are larger hurdles to overcome.
