Businesses expand and contract, following the whims of the economy, and when they are in a contraction phase, IT staffs are often required to help close remote offices. TechRepublic had to close four offices last year, and other companies are suffering the same misfortune because the economy is slow to recover from the malaise of the last two years.

From an IT perspective, the process can be divided between shutting down the remote office’s network and servers and shutting down the client side of the network as the staff disappears. TechRepublic looked at the client side of this issue in the Support Republic article “Blend tech with tact when closing a remote office.” In that article, TechRepublic IT director Troy Atwood and senior systems administrator Joe Zink described how they handled the closing of our ITRadar subsidiary in Minneapolis during the summer of 2001. In this article, Atwood and Zink discuss the server and network issues they confronted during that office closing.

More preparation, fewer boxes
Putting the personal and personnel considerations aside, the server issues of closing an office are still quite different from the client issues.

“It’s easy with a desktop, because you can put your fingers on it,” Zink said. The server and network side of the equation has a physical part to it, but it also has internal and external services added to the physical part of unplugging wires and moving boxes.

In fact, before you can even tackle the physical aspect, you have to inventory every office for services. You should already have an inventory of services at each location but you may need to augment that with more detail prior to an office closing.

When TechRepublic acquired ITRadar, Atwood had all the IT billing approvals transferred to him. Those bills included the office location, type of service, and circuit identities associated with them. This information was useful because it helped uncover redundant or unused services that could be eliminated. It was also invaluable during the process of closing the office.

Atwood said that shutting down a network at a remote office essentially involves these three areas:

  • Internal services
  • External services
  • Hardware

Here’s a look at what each aspect entails.

Remote internal services can hit home
Remote offices may be using two types of internal services—independent and distributed services—and just about any location will have both varieties. Independent services are things like local WINS or print servers that serve only the remote location and are not interdependent on servers in the organization. Distributed services are those that are closely tied to other servers throughout the organization, such as DNS and mail servers.

TechRepublic has segregated services to separate servers, so it is simple to keep track of which machine is doing what, Atwood said. That isn’t the case at some companies, particularly at remote offices where a limited number of servers may be providing multiple services. But even if it’s difficult to sort out which services a server is running, it’s essential to determine the effect of taking it down.

“You have to look at the apps you are running on a server,” Atwood said. “If I kill this server, will a [dependent] server cough up a lung in Germany?”

Some distributed services are particularly sensitive to what happens elsewhere. For instance, Microsoft Exchange Server reacts badly if it loses one of its distributed servers, Atwood said.

Before attempting to shut down an Exchange server in a remote office, you need to move its DNS and MX records and functions to your main office. ITRadar had its own primary DNS point when it was an independent company. But as part of the process of joining the two companies, ITRadar became a secondary DNS point for TechRepublic, and all of IT Radar’s Exchange mailboxes were migrated into TechRepublic’s Exchange environment.

When it came time to close the ITRadar office, the ITRadar server had to be removed from TechRepublic’s Exchange organization before the server in Minnesota could be shut down.

File services are kind of a cross between independent and distributed services. You could probably shut down a file server without upsetting other servers, but you might need the information for certain applications or user groups that exist in other locations. Zink said it’s best to make sure you check all enterprise applications that were used at the remote office for connections to applications at other offices, especially the main office.

External services can affect the budget
Although internal services are full of potential problems, at least they’re under your control. That is not the case with external services, which generally have contracts associated with them.

Common external services include:

  • Internet connectivity
  • Wide area network (WAN) connectivity
  • Telephone services

When ITRadar was closed, a fairly complex web of external services was at play. ITRadar had a satellite office in Iowa, which had a frame relay connection to the ITRadar office in Minneapolis as well as a cable modem for Internet access.

The Minneapolis ITRadar office had the frame relay connection to Iowa, a DSL connection for emergency backup purposes, and three T1 connections. One T1 was for Internet use, another was part of a WAN connection to TechRepublic, and the third T1 was a direct pipe to the back end of ITRadar’s Web site, which was hosted remotely. The Iowa and Minneapolis offices both had a number of telephone-related providers as well.

In an ideal situation, you would have 60 to 90 days notice before an office closed. That would allow you to contact data and voice circuit providers and cancel contracts. In many cases, moving PBX services needs to be contracted out so you don’t void the warranty. You also need to know if the PBX equipment is leased or owned.

As telephone services are being eliminated, Atwood and Zink recommend having all incoming calls for the closed office routed to one phone so that only one voice mailbox needs to be preserved.

The business will probably be stuck with a bill for a couple of months if the shutdown comes without warning. However, you should closely check all of the contract paperwork. Atwood did and found that the frame relay connection was expiring soon, so it cost TechRepublic nothing to end the service.

Pulling the plug
Once internal and external services are addressed, it is time to begin pulling plugs and boxing up equipment.

The servers will probably need to be shut down in a sequence determined by the services they are providing. For example, if the print server has nothing else on it, it will probably be shut down first. As other services are terminated, you’ll be able to unplug more servers.

While Atwood and Zink recommend having a packing company come in and box all the client machines, monitors, etc., they suggest you do your own boxing for the server room. There aren’t that many servers, and they—along with other network equipment—may be more fragile than desktop computers.

Once all the equipment from the closed remote office is moved to the main office, the servers and their previous functions will need to be reevaluated for reuse in the organization. In some cases, business objectives may require continuing a service for clients of the remote office, and you may need to bring one of the remote office’s servers back online from your main office, with some obvious tweaking.

Final word
In the end, knowing in detail what services the servers in the remote office are providing will make any changes there less disruptive on the rest of your network. In the case of TechRepublic and ITRadar, having a cooperative and professional local administrator at the remote office helped ease the transition and made the closing a fairly smooth process.

Have you ever had to pull the plug on a remote office?

Closing down a remote office is not an enjoyable experience, but it doesn’t have to be a painful one. If you’ve had to do it, did you encounter any surprises? What recommendations would you give? Post a comment in the discussion below.