During challenging economic times, with cut-throat global competition and declining margins, the notion of "doing more with less" has become a business mantra. Business leaders are constantly looking to minimise unnecessary costs, and are increasingly aware of the fact large proportions of IT budgets are spent on maintaining existing infrastructure rather than on business innovations.
In these circumstances, IT infrastructure budgets are heavily scrutinised, leaving IT leaders wondering how they will meet business expectations while still delivering on goals set by the organisation. Datacenter operations make up a large portion of total IT spending yet due to their business-critical nature are often overlooked when exploring potential cost savings. While there is no single "silver bullet" solution to these problems, several practical strategies can be implemented to reduce datacenter costs without having to defy the laws of physics.
IT leaders can become business heroes and make a real contribution to their organisations' profitability by implementing datacenter optimisiation best practices.
Servers have central roles in the datacenter running business applications, yet have considerable operating costs: they require rental of rack space, consume lots of power, and require external cooling in order to run effectively. Computing resources such as CPU, memory and storage are allocated to each server during installation. A typical physical server, on average, utilises less than 15 percent of these system resources. Not an efficient use of capital. And maintenance - for example, changing hardware configuration after the server has been provisioned - can be costly. Businesses can realise substantial cost savings by virtualising server workloads and consolidating them onto fewer physical servers with open source cloud technology such as OpenStack.
Even virtualised businesses can extract significant additional efficiencies. Those using aging servers to power their private cloud can significantly increase utilisation by refreshing their hardware. While large capital expenditure on new servers may seem counter-intuitive whilst trying to reduce spending, it can produce numerous optimisations and reduce operating costs, resulting in longer-term savings. For example, Intel's new Xeon E5 v2 processors contain technology to increase virtual machine density four-fold compared with typical four-year-old servers, allowing cloud computing operators to consolidate more applications onto fewer high-performing servers. This means reduced space in the datacenter, less power and cooling, and lower maintenance costs.
The growth of resource-intensive applications, high-volume data transactions and real-time technologies such as VoIP and video-on-demand are driving the need for greater networking bandwidth. Higher virtualisation density gained through server consolidation increases the demand for networking resources and leads to potential bottlenecks. Intel's next generation of 10Gbps network controllers provide significant performance optimisations in virtualised server environments such as increased bandwidth, decreased latency and lower CPU utilisation levels.
The burgeoning array of new enterprise applications and developments such as corporate data warehousing and "Big data" have multiplied the amount of digital data organisations need to store. And while storage has become more affordable, it still claims the lion's share of infrastructure capital expenditure.
As a critical component of any business, failure of a central storage system could have serious ramifications so infrastructure managers have to be careful when looking for savings in this area. Nevertheless, a number of techniques can be employed to manage storage growth more effectively. Data compression, which reduces the size of infrequently used files such as backups and disk images, can reduce the amount of storage required.
Data de-duplication is another useful option. This works by identifying duplicate files (or parts of files) on a disk and ensuring only one copy of the data is stored, providing considerable disk space savings. Openstack Swift allows cloud providers to extend the object storage system using middleware to implement inline compression and data de-duplication to optimise storage usage.
Implementing a combination of the above techniques will increase virtualisation density and storage efficiency in the datacenter. These strategies for datacenter infrastructure optimisation help companies reduce the pain of shrinking IT budgets while producing faster returns on capital investments through lower operating costs.