Virtualisation has long promised higher utilisation rates for storage, more effective distribution of network services and more efficient use of computing resources, and consequently reduced costs for those responsible for an IT infrastructure. While people have grown weary of the hype, two things have happened to give CIOs pause to revisit virtualisation technologies.
First, the demand for storage capacity has skyrocketed; IDC estimate that the amount of data storage purchased by an enterprise is growing at a rate of 80% per year. Given that current storage utilisation rates are only around 30% of capacity, the promise virtualisation gives of expanding that beyond 80% has a significant implication for the bottom line of any business. Secondly: the technology has improved. Industry leaders have developed products which make it possible to achieve these efficiencies in hardware terms, and also manage them so that the cost benefits can genuinely be achieved.
Virtualisation enables three main things: first, higher utilisation of storage capacity, second, simplified management and operation of resources and thirdly, the flexible and agile use of resources – for example, faster provisioning of services and reuse of resources as patterns of usage change.
The consequence of these is that overall costs can be reduced and essential services can be delivered more rapidly.
But virtualisation is not just about creating a virtual pool of disk capacity – there are many ways that virtualisation technology can provide benefits within a data centre.
Volume virtualisation, which has been provided for some years by storage vendors and virtualises the physical storage within the storage array. This capability has recently been extended to enable virtualisation over the networked storage infrastructure, allowing volumes to exist across arrays and to expand and contract as needed. This capability aids the delivery of a hierarchical storage environment, centralises management and operation of the entire set of storage resources and opens the possibility of a heterogeneous storage environment. A heterogeneous storage environment carries with it the benefits of flexibility in procurement of storage systems, to match business requirements, while still retaining a single management and operations mechanism.
On the other hand, using the concept of virtual Storage Area Networks (VSANs) allows a data centre to consolidate its storage networking resources into a single physical network from which separate, autonomous virtual SANs can be dynamically provisioned. This greatly improves utilisation rates of storage resources, provides a simpler storage network with reduced operational load, enables fast provisioning of resources and enables the simple delivery of an hierarchical storage environment, which is becoming of great interest with the advent of ILM processes.
There are many features of the network that can benefit from virtualisation. Physical connectivity virtualisation has been with us for many years, enabling many virtual LANs to be built from one physical network. Recently, virtualisation of intelligent services such as security and network monitoring resources, which are implemented in the network infrastructure, has extended the role of the network in delivering related benefits to the data centre.
Virtualisation of these resources allows them to be used at higher utilisation rates and managed in a centralised manner. This also enables the products that provide intelligent services to be placed at the core of the data centre network, providing services across all applications that may require them as and when they are needed. This enhances delivery of dynamic provisioning and flexibility of resource usage, resulting in an overall cost savings as less resources are required to service a larger overall network.
There are a number of areas where services are being delivered in the network infrastructure in a virtualised form. These can roughly be grouped as adaptive threat defence, including firewall services and intrusion detection, and application optimisation services, including server load-balancing and SSL offload services.
Computing power virtualisation is being provided in two models – one where a single physical server can offer multiple virtual servers and a second where multiple physical servers can be clustered to act as a single virtual server.
Again the main drive behind this technology is to improve utilisation rates of processors and to make more efficient use of processing power, to match the changing demands of applications at different times.
In this area, a networking infrastructure to link servers at very low latencies is starting to appear, based on gigabit ethernet, 10 gigabit ethernet or Infiniband. In some cases these infrastructures carry virtualisation capabilities that enable the automatic provisioning of computing power to virtual application servers. Such a server fabric switching environment aids the building of high performance clusters and utility (or on-demand) compute farms.
Virtualisation for the data centre
The promise of reduced costs and better resource allocation have potential to bear great fruit in the data centre. Implementing virtualisation technology in its various forms results in improved asset utilisation, simplified storage management and consolidated security and backup mechanisms. A bonus result is that business continuity programmes are more straightforward to establish, and overall business risk can be reduced – in addition to overall cost and productivity enhancements.
To that end, virtualisation technologies aren’t just a boon for those businesses looking out for their own data-centres, but are particularly key for any business outsourcing into industries for which compliance regulation has enforced a lower threshold of acceptable risk.
With technology now available that delivers these results in practise, as well as theory, virtualisation is coming of age – and is unquestionably the way forward.
Cisco Systems is exhibiting at Storage Expo 2006