Whilst IT departments may be under pressure to reduce overheads, it’s hard to get around the fact that a data centre consists of huge amounts of expensive equipment. All of which result in a significant amount of energy consumption and are therefore extremely costly to run. However, there are things that can be done to keep data centre costs to a minimum without compromising on performance:
Server virtualisation is a well understood technology that is employed in most, if not all, data centres. But virtualising networking functions is less commonplace. Virtualising capabilities such as firewalls and load balancers removes hardware footprints, thus reducing space used. It also reduces power and cooling requirements and hardware support costs. The benefit of this approach is that the software remains the same. And performance levels can easily be maintained through multi-way scaling. Whilst still gaining the flexibility that has long been enjoyed within the server world. This means network functions such as live migration and disaster recovery can easily be achieved without the need to replicate expensive dedicated hardware. And this keeps overall data centre costs to a minimum.
2. Multi-vendor approach
A single-vendor approach can often work out more expensive. It leads to reduced flexibility between platforms, tying a customer down further and allowing the vendor to become less competitive with their pricing and even innovation, without fear of the consequences. This lack of flexibility can also lead to lower performance. As the customer is forced to choose a solution from one particular vendor which may not be best suited to their needs. The reality is, vendors often excel in different areas. While one vendor may offer fantastic storage solutions, another may provide the best security solution. Sticking with just one vendor means you are unlikely to benefit from the best equipment across your network as a whole, whilst still paying more for it.
3. Seek advice on what you actually need
Just because a piece of equipment is the latest model with the widest ranging capabilities, doesn’t mean it is best suited to meet your company’s needs. That said, many new server and storage models offer significant energy savings which often offset the cost of the new equipment and can lead to significant savings compared with older models. All companies are different. So it’s important to assess exactly what you are trying to achieve. And what budget constraints you must adhere to along the way. If you are at all unsure, it makes sense to seek advice from the experts as it may save you money and prevent costly mistakes from being made.
A technical solutions partner with expertise across multiple vendors will be able to look at all legacy and emerging technologies. They will assess which of these can add greater value to your data centre, weighing up the costs and benefits along the way. Alternatively, if you already have a solution in mind, unbiased advice from a third-party technical engineer can help to answer any questions you may have. And confirm whether your chosen solution is the best option available. Or if a lower cost alternative would work just as well.
4. Ensure correct configuration
Incorrect configurations can have disastrous consequences. Causing bottlenecks and blockages which drain your network capacity, lowering network performance and costing you money. Similarly, as networks grow, new solutions and equipment can be “bolted on” to original systems as and when required, sometimes without any consistency or overarching plan in place. This organic growth almost always leads to inefficiencies and bottlenecks due to sub-optimal traffic paths and design. It is therefore important to ensure all new equipment is correctly planned, configured and installed. And your system is fully documented to allow for future planning to take place. If you are experiencing unusually slow or unreliable performance in your existing system it may be worth running a thorough examination of your network to identify any problems. It is good practice to periodically review your network systems in any case. This ensures they continue to deliver maximum performance for your business.
5. Consider refurbished
With manufacturers constantly developing newer, more powerful technologies, it can be tempting to think any new IT investment is best spent on the latest equipment. Of course, this is also likely to be the most expensive. However, the value of the latest capabilities becomes meaningless if they are capabilities you don’t need and are unlikely to use. If, in reality, your business does not need this level of performance, it may be worth exploring alternative, cheaper options to see if they can meet your needs just as effectively.
One of the alternatives to consider is refurbished kit. This may mean sending your own old kit for refurbishment to bring it back to a standard that is of value to you again. Or it may mean buying refurbished items from a reseller. Refurbished equipment often still comes with a warranty. And third-party support contracts will provide all the cover you need to maintain a fully functional system. Refurbished equipment often performs as new and can offer up to 90% cost savings so it’s well worth considering.
6. Use the cloud to manage your peaks
Provisioning for capacity peaks is an expensive proposition. Typically your data centre will have a base load and a “normal” operating maximum. But there will be burst of activity around peak times. You either have to provision hardware to cope with this or manage the demand downwards. Using hybrid cloud technologies coupled with Software Defined Networking techniques and Network Functions Virtualisation, the cloud can be treated as an elastic extension of your data centre. Dynamically being used to spin up capacity as required to manage those peaks. Traffic can be dynamically directed in and out of the cloud to manage load.
Most businesses that run a heavy compute load will experience these peaks and troughs. These will be dependent on seasonality, specific campaigns, product launches etc. Hybrid Cloud makes it possible to manage this. Paying for the capacity only when needed, rather than investing time and money in costly hardware builds which become under-utilised for the majority of the year.