Of course, these are average figures and I would encourage everyone to calculate their own cost of downtime.

In my experience, the figure is always a lot higher than first anticipated.

This makes the job of choosing fault-tolerant solutions much easier as you will be surprised at how low the cost of downtime can be to break even.

It then explores when a High Availability system should be considered vs. a Fault Tolerant solution.

All this wrapped up into a discussion around virtualisation and the benefits thereof, makes good reading.

Apart from the detailed information around IPMI, I was amazed that this guy is still running servers dating back to the late 1990s. Fourteen plus years is pretty good in anyone’s book…

Anyway, back to management of servers – I cannot comprehend anyone purchasing a system that cannot be remotely managed – that means remote console of course, mounting CDs, powering off/on etc. and having the ability to do everything to that system as though you were sitting in front of it.

All the major vendors have this lights-out or virtual-technician capability built in today. Make sure its implemented.

I would be interested in some comments about the ‘improving availability’ features of the new Windows 2012 operating system.

It looks great on paper even if it is overly complex. But my real question is whether it is real or mere smoke ‘n’ mirrors.

I am amused by a statement near the end of the article that says that the solutions “help to deliver” continuous availability.

I am not convinced that that is good enough. The only way to “actually” deliver is to use fault tolerant methodologies.

This can easily be remedied by fault tolerant hardware software solutions.

This high-level protection is getting easier and easier to implement.

Forty-five servers in a box sounds fantastic but there are going to be serious concerns over reliability.

One way to address this would be to double up the box45 concept and then run fault tolerant software between the two.

To me, that sounds like a fantastic idea and it will immediately negate the critics.

With today’s software, it is easy to do split site – even better!

As more and more services become virtualised, the pressure on the compute node running these vital services becomes ever more vital.

I use the word ‘vital’ from an availability and reliability point of view.

If you think you can simply re-deploy networking from dedicated routers and switches to standard computing environments (as per the emerging SDN developments) I don’t believe they will offer the reliability that is expected. ACCESS understands this, hence its partnership with the leading provider of fault tolerant infrastructures, Stratus.

A great opportunity to apply the belt and braces and re-deploy in some sort of fault-tolerant environment. It doesn’t matter whether it’s hardware or software, just buckle up.

In fact, loads of information.

Probably too much information.

Take the simple, worry-free computing approach and use either hardware or software to provide a fault tolerant solution.

It tackles the major benefits along with some misconceptions.

Have a read.

There is a new breed of fault tolerant server on the market that you need to be aware of and it’s really cheap.

They say that the most simple ideas reap the greatest benefits.

I can’t argue against that but I was most amused to find a claim by this company that their servers were fault tolerant (and yes, there is even mention of 100% uptime to tempt you) because of a unique dust filter that is fitted to the front and oh, a special Gold grade power supply.

Then I had a peek at where the company was based: Las Vegas.

What game are they playing?