In the past few weeks, not only did I move into a new home but I also managed the relocation of some critical servers from within our company to a new co-location facility.
To say I am frustrated and overwhelmed by DNS record updates, network wiring, routing tables and system latency is an understatement. Regardless of how much we plan for these moves, hidden glitches are always lurking -- waiting to attack the innocent IT manager trying to maintain some semblance of a productive work environment.
Laptop docking stations. The first problem we encountered involved a couple of executives on our team who have laptop computers with docking stations so they can keep the same system in the office, at home and on the road. Good in theory, but in reality when we changed the IP address of our main DNS server, the laptops didn't know anything had been altered.
Cries of "I can't connect to anything!" from laptop users set the tone for a frustrating 45-minute debugging session, before we recalled that the laptop in question was out of the office when all the DNS data was updated the previous Friday.
This highlights a problem with portable office stations: how do you keep track of whether they're in the facility and available during an important upgrade?
Meanwhile, at home. Attempts in my new home to connect via ISDN to the office with a new ISP have proven a challenge, too. First there was the three-week delay for line installation, but once installed, continued problems with the ISP gateway have stymied all attempts to establish a connection to the Internet.
Going through these hassles at the same time -- and starting to plan for the split of the office into two facilities -- reinforces the fact that even after all these years of engineering, computer networks often seem to be a huge house of cards.
Move something and the entire thing can tumble down, and if you want to change anything, don't be surprised if major problems ensue.
Useful, low cost alternatives to the NC
1997 was the Year of the Network Computer That Never Was, but in 1998 we may well see some companies offering low-cost, semiheadless systems for enterprise environments.
I've been in this industry long enough to know the history of client/server systems. I remember IBM terminals hooked to mainframes (which revolutionised the computer industry because tasks like editing were distributed between the client and the server).
Then X terminals were the next great hope -- a way to distribute the display task between the client and the server, but within a graphical environment: X Windows.
Now, as we're about to move boldly into a new millennium, the industry has again oscillated from having all the smarts on the desktop (PCs and Macintoshes) to having the smarts shared between the server and the desktop client machine. This time we're told it will be an identical personal computing experience, but with a dramatically lower cost per desktop. It's network computers to the rescue!
As I've said before, the idea of NCs makes a lot of sense in environments in which there's a full-time, high-speed network connection and an IS department that can manage the servers.
Previously, the industry made the big mistake of promoting NC systems as wonderful alternatives for expensive home computers -- a baffling idea.
If the main goal of deploying NC systems in a company is to lower the cost per desktop, there are some interesting alternatives worth exploring.
You can exploit these alternatives today to see cost savings, and they are built on proven and debugged technologies.
The Internet has created all sorts of busi- ness opportunities and one of the most interesting from an intranet perspective is smart com-panies retrofitting older computers -- both PC and Mac -- and offering them at discount prices.
Low cost is better
They aren't the sexiest, newest computers highlighted at Comdex or Macworld, but the point of modern computing is that you want to have low-cost systems on individual desktops, and fast, high-powered multiple-user servers. That's the basis of the entire network computer philosophy, but in reality we're already there. We can put low-cost systems on desktops with some creativity.