Acquisitions and mergers, rapid growth or simply poor management can each lead to uncontrolled development of IT systems to support changing business needs. Over time, this can result in unwieldy networks that are difficult and expensive to manage. When it comes to technology, less is often more — which is why it’s important to know how to reduce complexity and plan for growth.
A few years ago I worked for a company that acquired several small and midsize businesses. My job was to manage and integrate their IT systems. I was surprised by how complex and unruly the networks were, considering the functionality they provided was nothing out of the ordinary. Here’s a brief overview of two of those networks, which are examples of how not to do IT.
Company X was a small electronic-components manufacturer, based in one office with 30 employees. The company’s data was stored randomly across various devices: some on one of their two servers, some on a desktop (perhaps a server in a past life) located in the reception area and some on users’ individual machines. The systems were riddled with viruses, to the point where network throughput was reduced. A handful of outside salespeople were issued notebooks but used desktops when they worked in the office. Finally, the company’s Small Business Server was configured to use Internet Security and Acceleration Server (ISA), with a Cisco Private Internet Exchange (PIX) firewall between ISA and the ISP’s router. As is typical with many small businesses, Outlook PST files were used to archive e-mail.
Company Y was a telecom supplier with about 100 employees located in one office. Company Y’s in-house IT department had 12 servers performing a variety of undocumented functions. Soon after Company Y was acquired, we had problems with the Exchange Server. The mail-store database refused to mount because it reached its 16-gigabyte limit. Nobody thought to impose restrictions on mailboxes or monitor the size of the database. The database was restored from tape several times — taking four hours on each attempt — during a series of failed efforts to get it back online, and Company Y was without e-mail for two days.
The examples of companies X and Y are not as uncommon as you might think initally. In the case of Company X, which had been managed by a local IT support firm, putting out fires was normal and there was little consideration given to protecting the company’s data. The main issues Company X faced were unmanaged data storage, employees who needlessly used two machines and ineffective virus protection.
Company Y was suffering from a “boys with toys” mentality — the more toys, the better. The server room looked impressive, but you had to wonder why a company with only 100 employees needed 12 servers. With a severe case of server sprawl and only a small IT team, automated monitoring would have been essential to ensure a good level of service.
Through my experience with such companies, I’ve developed some best practices over the years that will help you better manage your operations:
Develop an action plan. After several weeks of observing a company and its day-to-day operations, you’ll have a fair idea of the main problems and some possible solutions. Start by drafting or modifying IT policy based on your observations. Get support from upper management for any changes you intend to implement, and document the intended result as part of IT policy. Ensure that any adjustments you intend to make fit the needs of the business, and that user acceptance is included as part of the process. Don’t proceed with your plans until you have support from all parties.
Every technology has its limitations. You should make sure that you and your users understand what the technology is designed to do and how to use it appropriately. For instance, is it necessary to deploy Exchange Server 2007 in a small branch office with 10 workers if there’s a reliable link to a hub site with appropriate capacity to support Outlook in cached mode?
Some steps to take when developing an action plan: Determine the most common problems and their causes by analyzing help-desk data; ensure that tasks and technologies are appropriately matched; identify where there is duplication of effort and streamline operations where appropriate; ensure that all business data is secure, can be backed up and restored in a timely fashion; aim to reduce the total cost of ownership for each device on the network; consider virtualization as an option to reduce server sprawl and get the maximum return on investment from the hardware; and create a rollback plan for any changes you intend
Work with the technology, not against it. Outlook PST files are generally not designed to be used over a network; as file-driven storage, they are easily corrupted. Although far from ideal, Exchange public folders can be used to archive e-mail, as can NTFS file shares. Research and collaboration tools such as Microsoft’s SharePoint Server and OneNote can be used to further reduce the burden on e-mail storage.
Clean up the mess. The real work starts only after the necessary documentation has been prepared and you’ve got buy-in from management. Use audit data to identify where the company’s data is stored. Don’t rely on users or support staff to provide you with this information. Even if you do get answers, you can’t be sure the information will be comprehensive.
Don’t move data from one location to another until you’ve created an image of the source machine, so that if data gets “lost” in the process, there’s a record of exactly what was located where before the move. Once data is moved to a server, make sure users have seamless access.
Unless you’re able to back up every device in the organization, configure file-level permissions and Group Policy to enforce data storage on file servers. Use Windows’ Offline Files feature to synchronize data locally for notebook users. You can then be sure that you know where all critical data is stored and that it is safeguarded by security and backup systems. Don’t rely on users to decide where business data should be saved.
Plan for growth. Once your house is in order, you’ll be eager to prevent it from descending into chaos again. Going through the process of documenting intended changes and involving all the stakeholders helps you avoid problems in the future. Don’t expedite solutions and quick fixes unnecessarily when responding to requests for additional functionality or capacity. Step back and consider the options carefully. Being the IT hero in the short run could result in increased costs and bind your company to bad practices that are hard to eradicate.
Russell Smith is an independent consultant based in the United Kingdom who specializes in Microsoft systems management.