With 2009 about to roll in, this is a good time to look at ways of making your IT organization better. Specific resolutions will be unique to every environment, but these goals make sense for just about any IT group.
The beginning of a new year provides a great opportunity to refine some of the finer IT practices and to engage new technologies to keep our collective juggernaut rolling. Here are some pointers to consider for the coming year.
Note: This information is also available as a PDF download.
#1: Get realistic about cross-training
Cross-training is a great concept, but how many IT shops actually have a fully cross-trained group? I am convinced that at this juncture in IT, it has become impossible to fully achieve. One strategy is to establish a primary and secondary resource and invest the time and training required to fully cross-train for each major discipline. One example would be a Microsoft Exchange environment: Sure, every administrator can make and remove accounts, but what do you do about Exchange log file, database, and message queue issues? In this situation, training a secondary staff member to be ready to step in for these tougher questions might be a realistic approach for achieving cross-training for all systems and platforms.
#2: Go the extra mile in virtualization
Take a step back and see where you are in regard to virtualization. Are there systems that can be virtualized — but little things are preventing it? This may be the time to resolve these issues, such as USB license keys or large storage, before the physical system’s failure causes a bigger issue. Where is the next virtual frontier for your environment? Virtualized networking with virtual IP addresses, virtual desktops, or possibly application virtualization may benefit your organization. In today’s environment, virtualization is more than just a way to save power in the data center.
#3: Push back when needed
It seems as though every IT organization is stretched to the limit in today’s economic climate. This includes budget pressure, cost control, staff burnout, staff resources, and of course the technology itself. With the pace of IT and the demands associated with it, it may be time to say, “No, we can’t provide that with the current resources.” Of course, there can be consequences with taking a stand like this, but there has to be a limit. One way to help solidify your position is to seek numbers from the large IT analyst groups like Gartner that quantify how many staff resources an organization of your size and type may contain. This third-party view can help the requestors and the funders get the picture that more needs to be provided or less needs to be requested with regard to what IT can and cannot do.
#4: Get going on 64-bit adoption
Given the current hardware environment, operating systems and core applications are offering 64-bit platforms now more than ever. Some of the initial growing pains that went along with 64-bit systems have been addressed, including driver issues, backups, and antivirus software that was slow to be fully compatible with 64-bit platforms. One compelling example of why it’s time to adopt 64-bit operating systems and applications is the forthcoming Windows Server 2008 R2 release, which will be available only in the x64 version. (However, the x86 Windows Server 2008 initial release does have x86 support.)
#5: Fix the rounded corners
We all are overworked, and we don’t have enough time to get everything done. One consequence of this workload is rounded corners in the small things that make IT work well. Have we documented everything fully that was implemented in 2008? Have we finished removing test servers, accounts, or programs that are no longer in use? Chances are, there are small details that need some attention to clean up the environment to improve the overall quality of IT.
#6: Design and implement new systems with built-in disaster recovery
There is no finer feeling than implementing a new solution that has built-in disaster recovery (DR), especially if it is across two data centers. This can be achieved using technologies such as a virtual IP address to load-balance traffic across two or more systems, virtualization, clustering, and other mechanisms. Unfortunately, there is no one-size-fits-all solution for DR, so a system may frequently have a combination of DR technologies to provide the overall solution. An example would be a Web-based service that resides on two virtual machines and is assigned a DNS name of a virtual IP address, which is load-balanced between the two Web servers that connect to a mirrored database.
#7: Automate the little things
Do we really need to jump through so many hoops to change a DNS entry? How about the effort it takes to page someone for assistance with a server? Now is a great time to invest in small automated mechanisms that can take away the burden of such tasks. The open source software space has a great offering of tools that can automate e-mail for certain events, provide paging services, and more. With some of the new technologies like Windows Server 2008 core edition, we may make an automated solution that is simply a .BAT file, Perl script, or a PowerShell script — that is okay in 2009!
#8: Knock out the old operating systems
Windows NT, Windows 2000 — they are out there. This may be the right time to get rid of them. But it is not always that simple. First of all, if these systems need to remain, they can probably be a virtual machine. Hardware from the pre-2003 era can easily be accommodated as a virtual machine in most circumstances. That, however, leads to the other symptom. If these older OSes are made a virtual machine, the pressure is really off to remove that platform. So take great care in deciding what is kept and what needs to go. For the platforms that need to be removed, determine what needs to happen to make the system go away.
This also applies to the old hardware out there: Take a walk through your data center and tell yourself, “It’s 2009. Is this equipment a risk point for me? If so, what needs to happen to knock it out?”
#9: Get involved in some sort of peer program
Chances are, other IT pros are out there doing what you are doing or are about to do. There is no better advice than hearing the unbiased answers concerning what others have done to address a problem or need that applies to you. Seek out a local group, such as the VMware Users Group (VMUG) program, or other user groups that collect like-minded people to talk about the issues relevant to the market today. In a way, it’s better to actually go to these types of events, if they are close and accessible. Simply joining a conference call or reading a forum doesn’t engage in the same fashion as a live event. Plus the snacks are usually better in person.
#10: Delegate as needed
For most environments, the number of systems and processes in which IT is involved has grown immensely in the past few years. In my experiences, there is a line between infrastructure teams and the rest of IT and the business as a whole. Not necessarily a barrier, but a dividing point, where responsibility exchanges from one group to another. For these situations, delegating control to those responsible (from the business or development team) should be the next step in freeing up infrastructure IT staff resources and letting the development, application, or business owners manage what they need and are capable of managing within the IT environment. This can include stopping Windows services, setting permissions on DNS entries for their systems (typically CNAMES), and light database query permissions. If we spend our day doing low-level tasks that can be assigned via permission to the persons responsible for the application, we’re not managing our time well enough.