Blog

Two simple ways to prevent unwanted IT systems downtime

Posted by .

Looking for two simple ways to prevent unscheduled computer network downtime? Easy! Fault tolerance and aggressive warranties for all your servers. This assumes that you are using Tier 1 equipment (i.e. HP, IBM / Lenovo, Dell, Cisco etc…). If you are using clones or “whiteboxes” for servers this may be difficult or even impossible so torch those clones and get yourself some name-brand servers now. That may be the most important step you take in arming yourself against downtime.

Remember 2 things…

But back to my two points… Point #1: Fault tolerance. Fault tolerance describes the ability to withstand server component failures without losing the entire server. Most commonly, you will see this in the form of redundant power supplies and hard drives. Given that these two components are the most mechanical (re: moving parts), they are therefore the most likely to fail and as such should be shored up by making them redundant. The good news is that 99.9% of all brand name servers have the ability to be equipped with redundant power supplies and hard drives.

Hardware Redundancy

What does that mean exactly (in English)? Well, simply stated, in the case of power supplies you install two power supplies instead of one and both are plugged into commercial power. In the event that one of them fails, the server remains functional thanks to the second power supply. Once the defective power supply is replaced – and this can in most cases be done while the server continues to run – this is called hot-swap capability – your level of redundancy is back to normal and you have incurred zero downtime. The same can be done with hard drives and although the process is somewhat more complicated, at the end of the day, you can survive the failure of one power supply and one (or more) hard drives and suffer no downtime as a result. Very Simple and very cost effective.

Point #2

The second simple way to prevent extended downtime is through aggressive warranty strategies. Again, this will likely require brand name server products. Essentially, almost all brand name servers come with a 3 year warranty and that warranty usually includes next business day on-site service. So if you have a component failure of any kind, the manufacturer is responsible for showing up on-site to replace the defective parts by next business day. All that is good and fine and a huge improvement over depot warranty (where you have to drop it off to be repaired), but what if the failure occurs late on a Thursday afternoon? After about 3:00 to 4:00 in the afternoon, server manufacturers tend to call this Friday meaning that they have until Monday to get you back up and running. So if it happens to fall on a week when the following Monday is a holiday, you may not be back in biz until sometime Tuesday; you could be down for upwards of 4 days and if this happens to be your mail server… yikes!

Warranty uplifts

Fortunately, Tier 1 manufacturers allow you to “uplift” that warranty to something much more aggressive such as 4 hour response (they have four hours with which to dispatch an on-site tech with the replacement parts, 24 x 7) or 6 hour call-to-repair (they have a guarantee to have your server back up and running within 6 hours of placing the support call irrespective of the failure, again 24 x 7). So, compare being down for 4 days to being down for 4 hours (or perhaps a little longer) and then do the math. Actually, I’ll help you with the math: the average $5M company burns ~$2k per hour in unplanned downtime and I recently spoke with a law firm with approximately 100 seats and they estimate a loss of almost $20k per hour of unplanned downtime!

In the final tally…

Do fault tolerance and aggressive warranties cost money? Of course, but not all that much and I will bet a week’s salary that it will be significantly less than the cost even one hour of downtime for 99.999% of all businesses that rely on their computer networks (isn’t that everyone?).

Just do it!

So if your network servers are not equipped with redundancies and aggressive warranties, ask yourself if your IT service provider has your best interest in mind and then talk to us. We insist on these simple inexpensive methods because we have seen the results of extended downtime and it ain’t pretty.