• Home
• Articles Archive
• Features List

• Media Pack
• Contact Us
• Subscribe

The growing cost of storage

From STORAGE Magazine Vol 6, Issue 4 - May 2006

Rapidly mounting costs are causing deep concerns for senior business executives as the need for storage hardware makes ever greater demands on their I.T. budgets. Ian Bond, consulting systems architect, Cisco Systems UK&I, looks at the implications.

In the increasingly distant past, 'data storage' meant hordes of filing cabinets and large, silent rooms to house them. Fortunately, the era of digital business allows us to keep our information safely in disk arrays that no one is likely to trip over. In this business environment, the management of storage has traditionally solely been an issue for the IT department, not the board or business owner. However, mounting costs are getting senior business executives more concerned with the demands storage hardware is making on their IT budgets.

The cost of storage, in absolute terms, has been dropping - massive production facilities in the Far East and elsewhere have brought us cheap hard drives, with storage costs down to as little as 25p per gigabyte. This is compared to closer to £1.80 per gigabyte only five years ago. To give a sense of how much impact this kind of cost reduction could make, Cisco itself manages more than 4 petabytes (4 million gigabytes) of data.

The reason costs have been mounting is that storage requirements have been growing at a rate that far outstrips these cost reductions: IDC estimate that the amount of data storage purchased by an enterprise is rocketing at a rate of 80% a year. As such, new storage facilities constitute a substantial proportion of overall IT spend. With the increased costs, business managers are finding they need to be thinking about their storage needs more strategically.

More and more storage
There are three major reasons why storage requirements are growing quite so dramatically. First, an increasing amount of data is being captured by applications and business processes. Crucial emails, documentation, recorded audio calls, etc. consume storage capacity insidiously. Cisco receives 10 million emails per day worldwide and counting - that is a trend that will only continue.

In addition, new technologies such as IP telephony allow easy recording of voice calls. For example, an IP-powered contact centre may audit a random proportion, or all, of the calls that its agents need to make - storing the call information digitally. That's a lot of data. In the future, as video becomes part of everyday business life and usage - for example, by the use of on-demand video conferencing services at the desk - even more information is likely to be held over long periods of time.
Secondly, compliance with account- ing and other regulations, such as Sarbanes Oxley, now requires that data is retained in a safe and recoverable manner, and for a longer period of time than had applied before. As a result, the risk of information loss must be reduced, meaning that all business critical data must be archived, in case part may need to be recovered, even years later. If any compliance environment demands that an audit trail is required, an organisation must provide for this and have the data available on request. A plethora of regulations exist to protect both businesses and customers and this is, perhaps, the key change in the last 10 years, in terms of data storage.

Thirdly, risk mitigation as part of a business continuity strategy also has a role in driving storage capacity increases. Over and above the regulatory environment to which many organisations need to react, they have their own assessment of the impact of events on their ability to continue running the business or services that they provide. These events can be either natural or man-made disasters and are obviously something that needs to be carefully considered, given the seeming increasing scale and effect of occurrences of both types in the past few years.

Protecting the information that is vital to enable the survival of any organ- isations’ delivery capability can lead to two, three, or more copies of data being held at dispersed storage sites. According to research from Sun Microsystems, 43% of companies that suffer large-scale data loss never reopen, and another 29% close within two years. Sometimes the increasing cost of storage has to be weighed up against the risks you mitigate by having redundancy built into your storage infrastructure.

Escaping the trap
There are some policies, processes and technologies that can be put in place to help deal with this issue. Like a human brain, we only use a small percentage of the resources we have available. Implementing new storage virtualisation technologies, which effectively pool storage resources and simplify storage management, can increase utilisation rates to 50% of available storage. Current usage rates are closer to 20%, meaning there's potential to double the resources available to you, without purchasing additional storage media.

Equally, implementing an 'Information Lifecycle Management' (ILM) strategy, and putting systems in place to support intelligent archiving of non-critical data, is crucial. These systems can help to lower storage costs by assigning non-critical data into different, cheaper, storage, eventually moving older data from disk systems down to very economic high-capacity device types, such as hybrid or pure tape backup systems.

How Can Virtualisation Help?
One of the key places that virtualisation can be implemented is at the connectivity level - the point at which the storage networking infrastructure enables interconnection between application servers and the storage systems that are the repository of the data utilised by these applications.

Without virtualisation of this Storage Area Network (SAN), storage capacity tends to be held in physically separate silos, often on a per application basis. These silos effectively isolate the storage capacity from anything but the single application for which it was initially provided, preventing the flexible allocation of capacity and leading to the very low levels of utilisation of disk storage by end user data that is often seen in our data centres today.

If the interconnecting SAN infrastructure is built as a virtualised environment that is one physical network to which all devices are connected, but out of which virtual storage networks can be dynamically configured, then the flexibility of allocation of resources is immediately and significantly improved. This leads to obvious improvements in the utilisation of the storage network resources, but also has a knock-on effect, in that the storage array capacity is now available to be flexibly allocated on a far wider scale, leading to much higher end data utilisation rates of this storage. An often overlooked second side-effect is that it also provides a simple means of recovering capacity that is no longer required by an application and reallocating where growth is needed.

Virtualisation of SANs also helps the delivery of ILM solutions. Different tiers of storage can be placed in different virtual SANs, and the movement of data between the tiers controlled by inter-virtual SAN routing techniques, aiding and simplifying operational control and management.

Maintaining the balance
CIOs managing their businesses' IT infrastructure are facing the continual struggle to maintain the balance between the need to deliver the new application capability that the ever-changing business environment requires and minimising the associated operational cost.

Having insufficient storage is unacceptable to modern businesses, but simply buying more and more storage media will only increase management costs and continue to place a heavy burden on resources, as storage requirements continue to grow exponentially. New technologies and policies, including virtualisation technologies and ILM, must be investigated before costs spiral out of control. ST