Lower TCO of IT Infrastructure

IT architecture
I am a Net Admin with a medium sized company running a mixed NT domain. I am fairly new to the industry and I am looking to lower our TCO of our network. I believe that we have way too many machines to support our users and applications. We have about 30 stations and we have 13 Windows Servers plus 6 Linux servers which are 2 db servers, 2 development servers, ftp server, archive server. The Windows boxes are as follows: DNS DNS Exchange Info store Exchange MTA SQL Test Application SQL Production Application BDC PDC aka File Print Server Misc Storage Application Server(app is rarely if not never used) Visual Source Safe Image storage which contains a 1.8 terabyte device Veritas Server w/ Robotic Library This seems like a definite overkill. How can I determine which servers can be combined. Your help is greatly appreciated.

Answer Wiki

Thanks. We'll let you know when a new response is added.

Looks like you have a lot to sort through. There’s a good article here on Techtarget called “ROI/TCO Fruit Salad”. I think it has some good ideas about how to move forward in trying to figure out exactly what to do to lower your true TCO.

Here’s the link:

Or see the article below:

ROI/TCO Fruit Salad

TCO (Total Cost of Ownership) is one of the industry’s most commonly used acronyms. Coupled with ROI (Return on Investment) these two key figures determine spending decisions for much of the equipment in use today, particularly with tight purse strings and budgets. As such, many of the companies selling products have gotten rather creative with these figures — leaving the CIO comparing apples to oranges, or even bananas. What is a person to do with fruit salad? This guide is designed to help.

Any TCO comparison should be based on actual IT budget items. This should include:

Equipment costs
Replacement parts
Power budget
Floor space
Downtime factors (such as mean time between failure, scheduled downtime)
Any changes necessary to implement or accommodate the expenditure.

Any ROI comparison should be based on tangible savings over existing systems directly associated with the total cost of ownership such as:

Life cycle of equipment
Expected savings over EXISTING solutions
Labor considerations (fewer employees, fewer contractors, etc.)
Productivity increases
Support increases or decreases
Training increases or decreases
Hardware expenditures and/or savings
Hours of operation
In looking at the TCO categories, obviously, the equipment cost is a finite number that equates to the expenditure on the equipment.

Maintenance should be figured over the lifetime of the system. Bear in mind that this may be an escalating cost as maintenance contracts generally increase in price by 3-5% over the lifetime of the equipment. It is a good idea during contract stage to negotiate a fixed maintenance contract if it is available. These costs should occur after any warranty period, provided that the warranty offers sufficient coverage.

Depreciation should be figured with a word of guidance from your CFO. Some companies use a longer depreciation cycle than others. For instance electronics can be 2-3 years, applications can be anywhere from 2-5 years and cabling generally falls between 7-12 years. If this is not a capital expenditure, but rather an expense, depreciation will not be applied.

Replacement parts are factored if it is necessary to have hot-spares on hand. Also, some equipment may fall out of warranty and/or a maintenance plan before it is scheduled for a dumpster. Even if the equipment comes with redundant power supplies and backplanes, for instance, you may want an extra interface card. Pay attention to the mean time between failures figures on the equipment to help determine when you will need spare parts. It is good practice to estimate 75% of that figure. Also pay close attention to any component that could cause you critical downtime waiting for a replacement.

Power budget is an often-overlooked figure. The amount of energy, heating and cooling and/or monitoring that is necessary for each device should be taken into account. In data centers, where cooling is critical and UPS’s or back up generators are in use, it is important to calculate loads to be sure that you can in fact supply power and cooling with current and expected capacities. The incremental costs for any upgrades to these systems will become part of the total cost of ownership.

Floor space is another hidden expense. Just as each employee has an overhead figure assigned for things such as taxes and desk space, equipment has the same overhead figure. In a data center where floor space quickly becomes prime real estate, this figure can be an important part of the calculation.

Training should be calculated not only for the cost of the class, but also for lost productivity, travel and other related expenses. Keep in mind also, that whoever you are training could get bitten by a black widow (that actually happened to me), or run over by a truck (that didn’t) and you will want to make sure that someone else is cross-trained to assist while they are out. Also calculate the cost of training your end-users and their lost productivity while they learn to use the new system.

Downtime factors can be a bit tricky. This should include things like the time it takes for upgrades, how often upgrades are necessary (boy that number sure has grown lately!) and any anticipated overtime for IT personnel to perform the upgrades. It is a good idea to estimate 75% of the mean time between failures and anticipate the time it would take to change out a component. Count downtime for all parties involved. If 100 people rely on a server, for instance, and you anticipate that it will be down for 5 hours during its lifecycle, this factor must include the weighted or burdened salaries for all 100 people as well as the cost of the IT staff or outside consultants to fix it. This one factor alone can be the deciding factor between two vendors.

Any other costs ? talk about a catchall! But this category includes things such as changing out components in other equipment for functionality, additional racks to mount the equipment, cabling changes, conversions and outside consultants to handle the changeover. Also include additional backup strategies, WAN links, etc. If you are going to run in parallel testing mode with an old system, time to audit those figures should also be factored in. Gartner estimates that companies underestimate labor by as much as 50%. You will want to pay particular attention to this figure. It is a good rule of thumb to take your labor estimate and up it by 25% to cover the unforeseen ? that Murphy sure was a smart guy!

A note about weighted or burdened salaries: When figuring hourly costs for consultants and inside personnel, the average salary is only part of the calculation. Companies have expenses such as insurance and taxes that also accompany a person’s salary. Your CFO should be able to give you the figure you need, but if not, estimate salary plus 15% for taxes only and salary plus 40% for taxes, insurance and overhead allocations such as the HR department which is generally allocated across a company.

Now that you are well versed in total cost of ownership let’s look at how that compares to a return on investment. This is really a calculation to see how long it will take the investment to “pay for itself” as a ratio of cost to savings. I saw a software company once that said its software would pay for itself in one month due to the elimination of personnel and increased productivity. What they didn’t take into account was that it was going to take a month to get it implemented and another month to fine-tune it to take into account such things like mobile users, etc. Add to that the increase in help desk calls when the application started installing patches on everyone’s system and the one-month turned to about six. So? let’s get real!

Life cycle of equipment is the expected lifespan of the solution as implemented. Any other savings must be offset by upgrade times, etc. which should be part of your total cost of ownership. If your existing equipment is likely to have to be replaced in one year, the replacement cost will become important in the calculations.

Hours of operation is an important figure. If you company is a 24×7 operation, your equipment will pay for itself faster than an 8-hour operation. But more importantly, if you have 100 users per shift saving 20 minutes on one shift, or 100 users per shift saving the same time but on 3 shifts, it is a much quicker (translated better) payback.

Expected savings over EXISTING solutions not above and beyond another product that you don’t have. Taking ROI figures straight from a vendor and not knowing what goes into it can be a dangerous pitfall. Vendors commonly do comparative studies as a sales tool. Not knowing your environment, obviously, they cannot supply you with a figure for your situation. They can, however, compare their product to a competitors based on industry available information. Be sure this applies to your environment. This should include maintenance of your old system, replacement parts, etc. that you will no longer need.

Labor considerations may be a negative number or a positive one. Be wary of predicting savings for an entire position. Chances are, the employee will start doing something else, particularly if they work in IT. Contractor savings can be a big number if you are continuously relying on outside resources. This should be all labor costs associated with your new system minus labor costs on the old system. Multiply this number by the number of shifts necessary. Be sure to figure in labor costs of upgrading your old system. Multiply this number by weighted salary figures and you will have an overall labor savings/cost figure.

Productivity increases should include things like decreased backup time, one hour less per day to do the same task than it took on an existing system, etc. Multiply this number by the number of shifts and number of employees necessary. For instance if you are installing a new accounting system, you will only factor in the people that will actually use it, not the total number of employees. If this is a new sales or customer relationship management and you expect each person in sales to increase their revenues by 2% because of a productivity gain, this is where that figure will appear.

Support increases or decreases figure your support costs on your existing system that are not included in the labor considerations or expected savings above. This could include things such as one less T-1 line, dropping some sort of support agreement not related to maintenance such as software support contracts. Be sure to figure in overtime labor if after hours support on your support contract is at a higher premium. This figure should include all hourly support agreements or support contracts that require a fee per call or flat fee.

Training increases or decreases should not include the training in the total cost of ownership, but things such as follow on user training. If it took you two days to train an end-user before and it will only take one now, the difference figure should go here. If the new version or upgrade of your existing platform was going to require 40 hours of training and this one will only require 8, again, the difference should go here. Multiply this number by the number of shifts and people necessary. Be sure to use weighted salary figures for people involved ? we are trying to stay apples to apples here.

Hardware expenditures and/or savings should include any new hardware you were going to need for your existing system. This could also be a hardware only cost for instance the difference between adding a 24-port switch and upgrading to a 48-port switch with better capabilities. It could be the savings of a few hard drives or additional tape drives, etc. For example, if you were going to upgrade your old PBX and it would have required an upgrade to the phones also, as you included hardware costs in your total cost of ownership, you should also reflect the savings of not having to upgrade your PBX hardware and phones here.

Now, simply take your total cost of ownership and divide it by your overall savings. If it is a $15,000 system and you will save $8,000 total over the same three-year period, then in 1.875 years, the system has paid for itself. By examining each of these topics, you are in, effect, comparing TCO for the new system against incremental or continued cost of ownership for maintaining your same system over the same period of time. More than likely your current system has already paid for itself. But no system is without upkeep and maintenance.

If you are looking at figures to compare two NEW systems for purchase, you can compare TCO to TCO on the systems, or ROI to ROI on each of the systems compared to your existing system. Apples to apples, oranges to oranges, and bananas to bananas are a lot easier to figure than fruit salad.


Carrie Higbie, Network Applications Market Manager, The Siemon Company
Carrie has been involved in the computing and networking industries for nearly 20 years. She has been involved in sales, executive management, and consulting on a wide variety of platforms and topologies. She has held Director and VP positions with fortune 500 companies and consulting firms. Carrie has taught classes for Novell, Microsoft, and Cisco certifications as well as CAD/CAE, networking and programming on a collegiate level. She has worked with manufacturing firms, medical institutions, casinos, healthcare providers, cable and wireless providers and a wide variety of other industries in both networking design/implementation, project management and software development for privately held consulting firms and most recently Network and Software Solutions.
Carrie currently works with The Siemon Company as the Network Applications Market Manager where her responsibilities include providing liaison services to electronic manufacturers to assure that there is harmony between the active electronics and existing and future cabling infrastructures. She participates with the IEEE, TIA and various consortiums for standards acceptance and works to further educate the end user community on the importance of a quality infrastructure. Carrie is one of the few that chose to work with applications and networks providing her with a full end-to-end understanding of business critical resources through all 7 layers of the OSI model. Carrie currently holds an RCDD/LAN Specialist from BICSI, MCNE from Novell and several other certifications.

Discuss This Question: 2  Replies

There was an error processing your information. Please try again later.
Thanks. We'll let you know when a new response is added.
Send me notifications when members answer or reply to this question.
  • Amigus
    That's a lot of servers! Without knowing all the details I'd suggest: Using Linux boxes for the DNS servers, and the Exchange MTA, doubling up the task of secondary name-server and sendmail/postix MTA on one box. Combine the Test and Production SQL servers unless there's a good reason not to do so, such as the "Test" application crashing the entire server. Combine the BDC, Misc storage, File/Print and even the Application server, onto one box, leaving the PDC to do it's job. That should actually spead up your network unless some users end up authenticating from the BDC for some reason. Combine the VSS with the Image storage if it makes sense given the two servers are probably similar and do similar but in some ways complimentary processing. If my math is right that brings you down to 6 servers from the original 13 given in your list. Regards
    0 pointsBadges:
  • Solutions1
    As noted by other respondents, there are a lt of dimensions to TCO. One perhaps not mentioned was to maintain "TCO over time." If, for example, you swap out your many servers for, say two or three 2005 vintage servers, you will be state of the art for 2005. However, to move ahead in, say, 2006 or 2007, you would then have to swap out big chunks of your infrastructure rather than in little chunks. Sometimes people who consolidate too far end up paralyzed, because to upgrade they have to put too much money or, worse still, too much organizational energy on the table. Therefore do a multi-year draft plan that leaves you maneuvabilty in the out years.
    0 pointsBadges:

Forgot Password

No problem! Submit your e-mail address below. We'll send you an e-mail containing your password.

Your password has been sent to:

To follow this tag...

There was an error processing your information. Please try again later.

Thanks! We'll email you when relevant content is added and updated.


Share this item with your network: