We understand everyone’s trying “to do more with less” given the economic climate, but it’s frequently the consultant’s responsibility to maximize clients’ technology investments and ensure systems and infrastructure efficiently power the client’s business operations. Sometimes that means having to do more with more.
Whether a client should upgrade or replace workstation computers is one thing; whether a client needs to upgrade or replace a server or servers is another issue altogether. License costs, migration and project planning expenses, and new hardware requirements all conspire to push server investments to seemingly precipitous heights. The simple (and refreshing) truth is that it’s not the independent IT consultant’s or consultancy’s fault that server upgrades or rollouts are expensive. It’s similar to the client being frustrated with an HVAC installer because the client lives in, say, Arizona where it’s very hot and the 15-year-old air conditioning system needs to be replaced at significant cost.
The tough and complex question IT consultants frequently face is: Does a server or group of servers need to be replaced, or can they be upgraded? Maybe a database server fails. Maybe a domain controller (DC) doubling as a file server consumes all free disk space. Or maybe a firm hits Exchange 2003’s 75 GB mail store limit. Do each of these scenarios require that the client purchase a new server, even if the current server is getting long in the tooth at four or so years old? Of course not.
Maybe the database server failed because it needs a new $249 power supply. Get the part, slap it in, and you’re done. Problem solved.
It’s possible the DC with no free space just needs a pair of additional 500 GB or 1 TB SATA drives installed, leveraging an existing onboard RAID 1 controller, to provide immediate additional disk space at low cost. No big deal.
But the answers aren’t always that simple and straightforward. Take the organization that’s reached Exchange 2003’s size limit and can’t (for whatever reason) archive mail or further reduce the mail store. The organization may need to move to Exchange 2010 (and the 64-bit server chassis required to make the leap), an open source solution, or something else. Either way, there’s going to be a price to pay. But if the client operates a 100- or 200-employee company that generates millions in revenue, in part by using email, then certainly they’ve budgeted funds to power such operations, right? If not, now’s the time to find the money because they just hit the current system’s capacity.
Those examples are cut-and-dried. The real world, however, deals in abstracts. Here are some guidelines consultancies can adopt to help clients maximize server investments:
- Research whether the existing server is covered by a manufacturer’s hardware warranty. Read the fine print, and if a failed server is still under warranty, hold the vendor’s feet to the fire. If a vendor originally sold four-hour on-site repair with the system, get the wheels moving. Often I’ve found my office’s engineers can diagnose an issue the vendor couldn’t, then we remind the vendor they’re obligated to send their own technician with the part within a specified period. This is especially true for equipment a previous IT provider may have sold and deployed for the client.
- Determine whether a hardware fix (such as replacing a failed hard disk, power supply, or similar failure) costing $500 or less likely means the server can be depended upon to provide another year of trouble-free and reliable service. If so, recommend it. If not, the client should consider replacing the server instead of repairing it. Subsequent downtime and repair costs resulting from other anticipated or age-related issues will exceed the costs of deploying a new server.
- Ask whether a downed server should even be repaired. If a new power supply is needed in a noncritical but necessary database server that’s suffering thermal issues and out of hard disk space, it’s time for a new box. Unless the server’s still covered by a manufacturer’s hardware warranty, the cost of repairs approaches the cost of a new machine.
- Consider migrating an older or failing server‘s roles and responsibilities to another existing server. Newer servers may have been deployed after the older system, thereby creating a possible opportunity to leverage existing capacity on the newer machine.
- Investigate virtualization as a solution. Organizations used to have to deploy separate servers to enable isolating memory or services for email, databases, applications, remote access, and other functions. Virtualization technologies offer organizations with multiple aging servers to consolidate backend services on fewer machines while increasing availability and fault tolerance.
There are exceptions to every rule, but these guidelines should help maximize clients’ IT investments and help minimize the sinking of additional needless costs into a machine that’s exceeding its service life.