The case for utility computing

With more and more companies demanding efficiency and flexibility from their storage, utility storage is one of the best options, argues Correy Voo, Head of Storage at BT.

With more and more companies demanding efficiency and flexibility from their storage requirements, utility storage is one of the best options, argues Correy Voo, Head of Storage at British Telecom.

Traditional storage models require that an organisation buy discrete storage systems for each of its sites or acquire equipment for its data center locations. This is a headache and a major cost for companies that are growing fast or are very distributed. Not only is the cost of purchasing dedicated, discrete storage systems, but the expense of locally managing, supporting and maintaining resilience in those systems may even be more significant.

To quell some of these costs, it may be time to revisit utility storage. Utility storage is delivered using a shared infrastructure model and typically only network connections to the storage provider's data center(s). This enables companies to cut costs in terms of outlay on equipment and demands little management on the ground, as this responsibility falls to the service provider. Moreover, as part of a huge infrastructure, often the more storage that is used, the cheaper per unit it becomes.

Avoiding Wasted Capacity

The different parties involved in the provision and procurement of storage often have conflicting views on the amount of space that an organisation requires. The expectation by most businesses is that storage needs will increase and expand. Installing new equipment and storage is costly therefore vendors will often encourage purchasers to make provision for their future potential storage needs, to ensure that they do not fall short and have to overhaul their systems quicker than anticipated.

Traditionally, the allocation of storage within a business is done in such as way that space is 'locked' away from new users or servers. This happens when users over estimate their requirements and add (sometimes significant) contingencies to their 'real' needs. However, this means that there is a high level of wasted capacity in very similar percentages to traditional direct-attached storage (DAS). This is inefficient and can generally be avoided by using on demand utility storage solutions.

Most companies cannot dictate how their business will grow or how their storage needs will evolve. Therefore it is never going to be accurate to try and 'guestimate' the requirements of storage in the future for a growing, changing enterprise. Utility storage offers flexibility to businesses, offering the opportunity to buy as and when storage is needed, so they are not paying for units of storage that they do not require at that moment in time.

Utility storage also allows organisations to buy extra or less storage during times of fluctuation in their marketplace. A retailer may need a significant increase in storage provision over the Christmas period but a sharp decline in that storage in January. Utility storage can be scaled up or down to respond very closely to a business's needs as the shared provisioning architectures allow for this flexibility across multiple customers.

Compliance Services

The amount of regulation around the legal use and retention of electronic information is increasing and so too are the penalties for non-compliance. Data protection, freedom of information, criminal laws and basic employee rights (including health and safety) are just a few of the issues facing businesses. To meet these responsibilities, good information and storage management is vital. Utility storage offers a cost effective hosted environment that can grow to accommodate increased demands and can be readily used to store less critical, longer term, data, thereby reducing the burden on a business' main storage facilities. Utility storage models can also provide much of the tiered storage capabilities required to support Information Lifecycle Management (ILM) methodologies.

Capacity-on-demand vs. Utility storage

Capacity-on-demand is another pay-as-you-go based storage model. However, unlike true utility storage, the majority of services on the market are, in reality, nothing more than disguised leasing models with deferred payments and potentially heavy commitments for usage. Customers are still required to specify up front how much storage they plan to use. If their actual consumption falls below expectations, they may be required to pay substantial penalties for non use. Customers should only pay for what they use, there should be no hidden costs.

Also, other storage provision methods do not have the flexibility of a true utility model. Typically, once storage is purchased, it cannot be given back to the vendor once it is no longer required. Basically, users cannot easily scale down, they can only scale up.

The latest and greatest storage

Storage systems are advancing rapidly both in terms of manageability and overall functionality. The problem with traditional storage models is that if a customer buys the best systems available today, it will most probably become an obsolete technology relatively soon after. However, as systems are purchased at a significant cost, organisations are unable to trade them in for the proverbial new model.

For utility service providers, updating the technology is core to the overall business proposition. As with any true managed service, the evolution of a technology platform can be dealt with by the service provider. By using utility storage, customers will be using systems at the cutting edge of technology. This is a very important factor for keeping ahead of competitors in today's fast moving environment and is a major selling point of service providers.

Moreover, with the onset of the digital networked economy, buying a system that will not be interoperable with different applications in the future could mean the storage provision will have to be scrapped because it can no longer work within a rapidly converging network and IT services environment.

Monitoring and charging for usage

As utility providers bill businesses for storage services they naturally have to monitor customer's usage with accuracy and transparency. Companies do not have to employ costly software and use valuable IT resource for this process. This monitoring capability is valuable to customers who may wish to implement internal utilisation charge mechanisms to share the burden of cost among the consumer users and organisational units.

Business continuity and security

Traditional storage systems are located on customer sites. If one of the customer's premises suffers interruption, this has the potential to bring part or even the whole of the storage network to a standstill.

As utility storage uses a centrally hosted, shared infrastructure, when customers suffer interruptions, system backups allow customers to continue to use systems from the service provider's data center. Reconnection to the storage from appropriate alternative business locations will enable customers to keep working.

Service providers are expected to provide duplicate infrastructure components, with automatic fall-over to ensure faults on their systems have minimal or no impact on customer service availability. Services are protected by managed security and also proactively monitored to identify and resolve potential problems on a round the clock basis. This means business as usual for customers in most eventualities.

Unlike utility storage, traditional storage provision and capacity-on-demand storage is often based on customer sites. Data stored in the capacity-on-demand model is still subject to the same threats of interruption as those of traditional storage models.

The bottom line

Utility storage can offer customers a system that is tailored to their specific data requirements and business model.

However, it is important when businesses look for storage provision that they check the fine print. They should ascertain whether the service offered is on a shared infrastructure thereby avoiding significant outlay costs on hardware and installation. Customers should also find out how well their data will be protected from threats and interruptions and how service providers will prevent other users obtaining access to their data.

Utility storage models should always be transparent and customers should only pay for what they use. There should be no hidden costs and businesses should check that there are no penalties for non-usage (as they would with the capacity on demand model).

Finally, digital storage purchasers should subject their storage providers to the same rigorous evaluation as their data will be subject to, because failure to do so may result in increased costs, compliance penalties, business interruptions and a significant dent in their bottom line.

For more information:

Utility computing

Utility computing to bring benefits, for a price

Finding utilitarian uses for utility computing

About the author

Correy Voo is head of business technology solutions and a principal IT consultant for the Information Management and Infrastructure Solutions unit within BT's Global Services division. He specializes in systems integration, information management, business & process management and information security.

Dig Deeper on Storage vendors

Start the conversation

Send me notifications when other members comment.

Please create a username to comment.