Expanding on what Worthy Oracle Keith said, one of the things that I am hoping to bring to CRT is my experience in the front lines of Broker/Agent technology. Working at Prudential - Chicago almost from its birth, I witnessed and implemented 3 incarnations of their network and web services. Through this, the most challenging was maintaining software and hardware that met the needs of the users, but at the same time didn't break the IT budget.
One of the reasons that the hardware/software was so challenging, was that it wasn't setup like a traditional company, where each user has their dedicated machine, and upgrades can deploy departmentally. Instead, each office had several shared workstations which dozens of agents would pass through daily. Each user works differently, has access to different software and web services on their personal machines, and relies on the branch machines to play nicely with whatever they may be working on. In other words introducing an unknown amount of variables for software and usage requirements - the greatest fear of any person in technology, the random factor.
This lends itself to an interesting topic, as 2007 is going to bring the simultaneous launch of both Microsoft Vista and Office 2007. Although Prudential had been budgeting for this upgrade for the past year and half, they had been preparing for the upgrade and getting all the pieces in place for close to 3 years prior. The key to understanding your upgrade path is to understand the business impact both in terms of financial and usability. Each companies mileage may vary with the same solution, so these factors need to be fully identified and understood.
It is often debated if there really has been any substantial change to MS Office since version 2000 which would warrant the cost outlay for an upgrade (there have been 2 version released XP and 2003). Many of the new features introduced rarely affect your average user, as they are power users. However, as new machines enter your network or colleagues begin to use the latest version of an application, you immediately lend yourself to version conflict, frustration that only some machines can work with certain programs/files, and an overall feeling of the urgency to upgrade. STOP right there, as this feeling is out of panic, as if you take a step back and look at the costs, time, and timeline of upgrading, you quickly realize that there has to be an alternative route.
Understanding the rate of new software and how it changes your business workflow is the first step. In Microsoft's case, their goal is a less than a three year life cycle of its main products (Windows/Office), which is why they have introduced Software Assurance . However, each person has thier own opinion as to if version 'X' is that much better than the version they are using to warrant an upgrade cost, and lets us not forget that they slightly missed the 3 year mark since XP was released in 2002. Let us now apply this to the impending 2007 launch of these new products.
Per workstation (including machine/cpu) the estimated cost for Windows Vista/Office 2007 is roughly between $3,250 and $5,000 (which is highly dependent on purchasing licenses at time of machine, purchasing license agreements and/or Software Assurance for bulk usage as well as will your machine even handle the newest version of Windows). In a company which has 100 workstations, that would be quite a bit of expenditure, as well as the time to rollout 100 machines would significantly shuffle the IT staff timeline. Another option would be to set a cycle for obsolescence for every 3-4 years therefore reducing yearly costs. However, with this, you lend a considerable amount of overhead to your IT staff as they are supporting 2-3 different versions of software, machines, and general configuration variance. In addition, you'll have machines that users just stay away from because they are the 'leftovers' or the ones that the software doesn't work with 'any of the other computers in the world'.
The strategy that was deployed at Prudential was two pronged. First, it consisted of staying with a core version of MS Office and only upgrading if absolutely necessary or 'bite the bullet' after two version releases (which in theory should be 4-5 years). Now one wonders what happens if a user needs the latest version to read a file created or some other scenario. Well, in dire circumstances, special use machines could be deployed to act as a stop gaps between the present and upgrade path. To lessen these special circumstances, a remote application server using 2X or similar could be deployed . This allows for a centrally managed set of distributed software and reduces the number of special machines. Therefore lowering configuration discrepancies and IT support time.
As always, there are other alternatives, one of which is the excellent and free OpenOffice.org. I strongly recommend every reader that is pondering the benefits of this article, to thoroughly investigate OpenOffice.org, keeping in mind what I mentioned earlier - with each alternative solution YMMV.
The second piece to this two prong attack is that Prudential moved the entire hardware/software architecture over to thin client computing, where each computer is controlled by a central server running a copy of Windows Server . The client machine can be a specialized hardware unit such as WYSE or Neoware, or it can be a software based client (such as Thinstation , 2X ThinClient or Diet-PC) that is run on commodity hardware. The pure hardware solution definitely takes some of the variability out of deployment, but adds the added (nominal) cost to deployment. The benefit of running a software based client is that you can either reuse otherwise obsolescent hardware or purchase extremely inexpensive refurbished hardware and run it all into the ground. The software client does have a learning curve, and can have compatibility conflicts with hardware/software. The final direction at Prudential was to use a hybrid approach; dedicated hardware clients for some departments and commodity hardware based software clients for the rest of the workstations.
This approach would therefore reduce your upgrade costs by approximately $1000-1500 per machine every upgrade cycle since the hardware is reusable. In addition, there is the cost savings for the operating system, since only one copy of Windows Server controls all the client machines, while keeping mind there is a per client license cost, which is considerably less expensive than a full copy of Windows. The cost for application software is a wash per machine, however you do extend your software purchases from every 2-3 years, to every 4-5 years which will lower cost extended over time.
As a rough estimate, the cost per machine would be $1000-$2000 which is considerably less than the estimated cost for a full upgrade with new hardware and operating system per machine. In addition, there is a savings of IT staff time savings for reduced machine deployment, troubleshooting (both due to centrally managed and less version/configuration discrepancies) as well as more time allocated for projects moving the company technology forward as opposed to maintenance.
In closing, as with all software discussions, especially those that develop unique strategies for balancing work efficiency and budgets - consult a Information Technologist well versed in software licensing. As some other wise IT Shiva once said, if it seems to good to be true, it probably is.
0 Responses to “Cost Wrangling the Random Factor”