VARs and MSPs can protect small business customers against server failures by holding a cold spare server for...
Server hardware fails -- happily, not nearly as often as it did 10 years ago, but it still happens. The technical problem with the cold spare server approach is that commodity servers aren’t necessarily interchangeable. Once an operating system is installed, it is closely coupled with the server, so swapping for a different server is hard. But if there is a common hardware platform, for one customer or across a number of customers, it can be cost effective to hold on to a cold spare server in case of a failure.
We use redundancy to protect against disk, fan and power supply failure. What about CPU, RAM and motherboard failure? Having a standard hardware platform makes it much easier to hold a cold spare server, and it allows rapid recovery from these non-redundant parts failing. If every server has the same chassis, CPU count and amount of RAM, then a single-server chassis can provide rapid recovery for a number of servers. It also minimizes the cost of holding spare fans and power supplies, as you will use the same ones across servers.
Setting up a cold spare server
The cold spare server is configured and cabled the same as the production servers but has no disks. If there is a non-disk server failure, then the disks are moved over into the spare and it is powered on, assuming the identity of the failed server. In one example, a customer had three very similar servers and a fourth chassis as a cold spare server. When a project required a new server urgently, the customer only needed disks for the spare to become a production server. Ordering a new spare chassis returned the customer to its usual state of preparedness for server failure.
To provide this availability for smaller business customers, solution providers can hold onto the cold spare server. It may not be dedicated to one customer, but shared across several customers who each pay a portion of the spare cost as part of the support contract and share the protection.
Shared cold spare server availability
If you operate in a smaller city, having a cold spare server may make immediate sense, because it can take a day or more for a warranty replacement to arrive. Getting customers operational the same day is a clear benefit. Having a spare may allow you to sell the customer on a lower warranty level with next-business-day response (rather than four-hour response), and then use your local spare to upgrade the service level.
Naturally you will need to choose the chassis carefully. You need to be sure you will sell a number of the same chassis before it is superseded; this will make the spare chassis cost effective. You will want to select a chassis that has just been launched and has a long supported life (so, one from a tier-one vendor). It is also important to select a chassis and configuration that will suit a large number of customers and uses within customers, which will make it easier to sell it enough times to make money. The sales team also needs to get behind the standard platform, proposing the standard platform rather than customizing an offering for each sale.
Another benefit of the standardization will be greater familiarity with that model of server for your engineers, speeding time to recovery when customers have issues. Being tied to one vendor will also allow you to invest in vendor certification for you engineers, further differentiating your team from competitors.
Committing to a specific chassis and configuration may also allow you to hold stock in advance of orders, enabling projects to proceed before fulfilment. But make sure that your server vendor starts the warranty clock when you install the hardware for the customer, not when you receive the hardware shipment.
Having a standardized hardware platform has a number of benefits for a reseller and for customers. The challenge is promoting a full standardization culture from sales to delivery and support.
About the author
Alastair Cooke is a freelance trainer, consultant and blogger specializing in server and desktop virtualization. Known in Australia and New Zealand for the APAC virtualization podcast and regional community events, Cooke was awarded VMware’s vExpert status for his 2010 efforts. Follow him on Twitter @DemitasseNZ.