The recent news reports
that I have read regarding organizations, both public and private, who were
impacted by the Zotob and Rbot worms is disconcerting. To the uninitiated,
those who were affected did not practice good patch management.
However, those in the
trenches know that keeping up with patches, especially at the desktop level, is
a daunting task. On top of that, patch management is just a small part of total
desktop management. Application installation, software updates, hardware
maintenance, training, security, and more are all part of desktop management.
In 1996, Gartner Research announced
the average Windows 95 desktop cost $10,000 a year to own. This includes,
besides the activities mentioned above, the direct costs of user support, lost
productivity, downtime, and administrative costs including depreciation, and
finance charges.
Some would argue even then
that the total cost of ownership (TCO) computed by Gartner was either
overinflated or underrepresented. Whatever the cost is today, (based on your
own TCO) there is no arguing that managing the desktop takes up a significant
amount of an IT department’s total resources.
The workload to manage
desktops is such that a whole market of desktop management tools have sprung up
to help us “control” them. Novell Zenworks, Intel Landesk, Hewlett-
Packard OpenView, IBM’s Tivoli TME10, or Microsoft’s Zero Administration Kit
are just a few examples. And even with these tools, the TCO for a PC just seems
to stay the same or is even increasing.
So given all this, don’t
you have to wonder if it’s worth it?
I know I did a few years
ago. I looked at my organization’s IT budget and the amount that was being used
to purchase and support PCs and said “there has to be a better way”.
That better way, after
some significant research and testing was a hybrid solution consisting of thin
client technology, Citrix, and Linux that would be phased in over time.
In a nutshell, the plan
was to provide a “desktop” to the end user via their browser and run
all their applications either directly from a Web server or from a Citrix
server. Their individual machines would have their OS replaced by a very thin
build of Linux and all machines purchased (new or as replacements) would be
Linux-based thin client machines.
This desktop solution, in
my opinion, was more secure and less susceptible to end user
“intervention”, virus breakouts, and emergency patches; thin clients
were easier to install and trouble shoot, and workers were no longer
“tied” to their workstation since they could get their own tailored
desktop from any machine in the organization.
In order to succeed, we
had to do two critical things. Get buy-in from the organization and
make sure we had a solid network infrastructure that had a very low
latency.
The buy-in began with the IT
governance committee. Fortunately, we had a very astute committee that, after
seeing the research and the solution in action, quickly bought in to the idea. The
next step was to eat our own dog food. The IT department made the move to this
solution. From there, it was time to woo top management. We knew that if they didn’t
understand what was going on, the plan would eventually fail. Again, at the
time I attempted this plan (and probably why I was comfortable in doing it) we
had an incredibly sharp and IT-friendly administration. From the CEO to the
CFO, they were on board and active supporters.
Meanwhile the
infrastructure work was going on, and we were revamping and expanding what was a
tired network to start with. So we began putting the solution into place,
department by department, and you know what? It worked great! Yes, we initially
had some kinks, but once they were worked out the network was greatly improved.
Unfortunately, this story
has an incomplete ending. While doing the project, our community voted to merge
local governments. So when the time came for the merger, we were not quite
finished. Overnight, the whole environment that had been primed and ready
literally disappeared. That ended the solution.
The point of this story
though is that I know the solution works and you can be rid of a great deal of
desktop headaches by ridding yourself of a fat client machine. I have seen the
beginnings of it, I know what the TCO was turning out to be, and if I ever get
the opportunity again I will seek to implement a similar solution.
And yes I realize that
this solution doesn’t fit 100% of the users but it works for the vast majority.
And I would rather be managing a very small subset of needy fat client machines
than an entire organization’s.
Its a bold move, switching from fat client
machines to thin client and delivering an organization’s desktop through a Web
browser. But with the proper planning and execution it can be done. Just make
sure you start at the beginning of
your CEO’s next term in office.
Keep up with the issues and challenges that uniquely affect
public-sector IT with TechRepublic’s free Government IT newsletter,
delivered each Tuesday. Automatically sign up today!