If you own a car, you know regular tune-ups are a must for avoiding unnecessary repairs. You also know that once you near 100K miles, you better replace the timing belt — even though it’s still working — or you will soon come to a screeching halt.
The same can be said for data centers; improper maintenance will inevitably lead to failure and delaying aged asset replacement could result in downtime.
No. 1 – Maintenance
Car maintenance basics include oil changes and tire rotation. Recommended oil change cycles used to be every 3,000 miles, but today many vehicles far surpass this — some BMWs can go up to 10,000 miles.
Tire rotation can be done every time the oil is changed, but maybe every 3,000 miles is too soon, yet 10,000 miles would be too long. The schedule will depend on the manufacturers’ recommendations and on how much you drive.
In data centers, there are also manufacturer recommended cycles. The oil in generators needs to be changed at certain intervals, plus, UPS’, batteries, capacitors and air conditioning filters need maintenance.
Preventative maintenance serves two primary purposes. 1) it reduces the risk of unplanned failures and; 2) it ensures that equipment is operating at peak efficiency.
As noted in a previous blog, the best data center maintenance program consists of preventive and predictive (condition-based) maintenance. You want to avoid reactive maintenance as much as possible since it results from something going wrong.
Preventative maintenance is standardized based on OEM suggestions, but it doesn’t necessarily reflect conditions like predictive maintenance does. Predictive or “condition-based” maintenance is becoming the norm with the growth of sophisticated monitoring and automation tools.
No. 2 – Replacement
No matter how much maintenance you do on your car or in your data center, components still have to be replaced after enough time and use.
In a car, tires may be certified for up to 60,000 miles and you could probably go to 70,000, but the longer you wait, the greater the risk of a blowout — which could cause collateral damage.
Similarly, delaying replacing aged assets, like UPS’, in a data center introduces operational risk and could result in a catastrophic failure. Not to mention, capital outlay is also compounded when asset replacement is delayed.
For example, putting off replacing $1 million worth of equipment for 5 years could quickly turn into $5 million because then, you are faced with replacing other assets that have aged over that period of time. Your risk profile is now 5 years greater and your expense exposure is 5 times as great.
This creates a bubble that is sure to burst. Optimizing reliability and cost means staying on top of your maintenance and asset replacements, not playing catch up.
No. 3 – Leasing
If you don’t want to be responsible for maintenance, repairs and capital replacements, you could lease a car, but go in with eyes wide open — the longer-term costs associated with leasing do not always make for the best value.
The same could be said for utilizing colocation. Make sure you are well informed and understand the total cost before making the decision. For example, consider what happens after the contract runs out. With car leases, you have the option to buy at the end of the agreement, but with colocation you need to think about what you are going to do at the end of the term – you can’t just drive away with your data center.
Could you stretch out the recommended maintenance cycles on your data center equipment — the tools that keep it up and running? Perhaps; but there’s no telling exactly what will happen. Do you really want to be the first to find out?
Can you delay replacing aged assets? Sure; but do you want to risk being the cause of business going down?