Linux, UNIX, and Windows: TCO Revisited

base_chakra writes "In response to Steve Ballmer's recent, vague criticisms of Linux and UNIX vendors' product support structures and pricing schemes, I've compiled a reference chart which might help others to compare licensing and support options available from several big-name vendors. Kudos to Red Hat for having the simplest, most accessible licensing and service plan of the pack--as well as one of the most affordable!"

Comments

My personal case study.

[consolidatedbord]

I've only been working in the IT field for about 1.5 years now, but in that time I have never, ever had to call Redhat support. (Everything Just Works(tm))

I have spent about 2 hours or less on the phone with Sun support.

I have spent a whopping 20+ hours alone on the phone with Microsoft support.

Unfortunately here for Microsoft they can't use the argument that it's because they have a dominating market and the ratio of support calls to them is therefore higher. Not in my client's server rooms. The ratio of the MS to "other" operating systems installed on the servers is dead near 50/50.

Just my personal experience.

The points made taken with a Grain(or pound)

[da_Den_man]

Of S.A.L.T.

To quote briefly from the points made:

n early 2004, Forrester conducted in-depth discussions with 14 companies that had been running Linux platforms for longer than one year to see what the costs really were. Several key themes emerged:

* Few companies know what they're really spending. Only five of the 14 kept detailed metrics - and each of those five found Linux more expensive (5% to 20%) than their current Microsoft environments.

Basically, this is stating that regardless of the figures any company gives, they cannot and have not kept accurate records of the TCO for anything. This reason is moot at this point due to the inept accounting practices.

* Preparation and planning activities took 5% to 25% longer for Linux than Windows.

Linux has not been in the marketplace (Business) so naturally planning a move to A COMPLETELY DIFFERENT method is going to require more planning and better preparedness. In my experience, this turns into meetings and more meetings, without the ability to truly resolve anything. Naturally, when you decide to shift PLATFORMS to something like Linux from Windows, there SHOULD be more planning and careful consideration of the differing methodology. Ask these companies how long it took to implement a datastore, or a full blown Notes / Exchange server, and I am sure the amounts of time/personnel were similar. In this planning stage, a company would want to make sure that there will be several contingencies in place to handle unknown elements.

* Training for IT employees was significantly higher for Linux than for Windows - on average, 15% more expensive. The reasons: training materials were less readily available, and customers spent more on training to compensate for the lack of internal knowledge about Linux.

This just means that more people need to be trained in how to best utilize the new platform. Again, this would apply to any new SOFTWARE application...which could be why there are such issues with upgrades and new installs now. The company usually views training as a last effort and minimal cost is spent. "We can get 'Joe' trained and he can train others" which can usually lead to misinformation and bad training, especially if 'Joe' has some other job to go along with the new one of Instructor.

* All 14 companies said it was difficult finding qualified Linux personnel in the marketplace to support their Linux projects. When they did find third-party help, they had less leverage negotiating hourly rates than with Windows consulting resources.

See the above reply to #3. There needs to be more people in the field with the ability to actually DO the work. This issue then falls into the fact that the bottom line for any company these days is "How much are we spending" instead of "How productive are we going to be". The leverage they discuss is geared towards how cheap can we get this done. Without developers to play against each other (and whether they can outsource later...which to my knowledge they really cannot with Linux) the company ends up paying the "going rate" for a decent programmer. Tsk Tsk (sarcasm intended)

IMHO, it seems that companies polled were looking for INSTANT savings as opposed to overall TCO. Not taking into consideration the savings at the next version of Windows (Not needing to spend XXXXXX dollars on an upgrade platform). Nor taking into consideration the fact that if there IS a problem with the implementation that someone can FIX it then, and not have to call the MFR for a patch...or wait for a patch....or hope for a patch....

Just MS spewing forth as normal I guess. Sad to say, but it always falls to "How much do we save the shareholders this week" rather than "How much do we save Overall?"