Plugging the storage gap

Storage devices are not buried away, they are on the network and make even archived data accessible. The Computer Weekly/ Posetiv...

Storage devices are not buried away, they are on the network and make even archived data accessible. The Computer Weekly/ Posetiv survey finds users are prioritising software that manages stored data. Antony Savvas reports

Two-thirds of UK organisations put storage in their top five IT priorities, according to the annual Computer Weekly/Posetiv survey of data storage use, published this week. The survey, which questioned 800 senior IT professionals at companies of various sizes, exposes the strategic nature of storage requirements in business.

The results of the study illustrate how companies are trying to cope with the continuing data explosion, as security threats and legislative moves by governments increase pressure on organisations to make sure their data banks are in order.

Investment in storage products and services is driven, the survey finds, by three key requirements: 

  • Information availability (53% of respondents)
  • Increasing network reliability (45%) 
  • Faster information extraction (39%).

These requirements are all network-related as they correspond to massive increases in data traffic from the greater use of e-mail, general office applications, enterprise applications and downloaded files.

Once again, e-mail is cited in the survey as the main culprit for generating increased storage requirements, with 60% of respondents saying e-mail takes up most of their storage space. This is down from the almost 70% who put e-mail in the frame in 2003.

Even with this fall in the 2004 survey, e-mail is still well ahead of general office applications (48%), enterprise applications (40%) and downloaded files (17%). And in the future, e-mail will remain a major headache, with 63% of respondents saying it will be the biggest drain on storage requirements. However, data storage for enterprise applications, at 48%, is growing fast, overtaking general office applications (47%). Downloaded files will rise to 20% in the next 12 months.

The biggest challenges companies face in storage planning, now and in the next 12 months, are:

  • Predicting storage growth (47%)
  • Security (45%)
  • Total cost of ownership (30%) 
  • Coping with storage management complexity (27%).

So which technologies are organisations relying on to store their data? The advent of storage networks has allowed firms to share data among a greater number of end-users and set up linked and remote back-up sites.

Last year, 15% of respondents said their organisations operated a storage area network and 17% said they had a network-attached storage system. Now 26% of respondents use Sans and 33% have installed NAS systems.

Further growth in networked storage is predicted for the future. In the next 12 months, 26% of respondents say their firm will invest in a San, and 31% are planning to buy a NAS system. Last year 13% forecast that they would be buying a San, and 12% a NAS package.

Susan Clarke, a senior research analyst at Butler Group, says, "The San has certainly progressed beyond the early-adopter stage, but it is not necessarily enterprise-wide - there are still a lot of small implementations. NAS is also well-established and is employed alongside a San in many organisations."

But it does not stop at Sans and NAS. A newer - and potentially better - technology is iSCSI, which has a number of advantages over the existing technologies used in datacentres.

The standards for iSCSI were only finalised about a year ago, but the Computer Weekly/ Posetiv survey finds that 7% of organisations are already using iSCSi and a further 7% say they will be using it in the next 12 months.

"The low take-up of iSCSI is due in part to the fact that suppliers have only really started supporting it in the past 10 months following the release of Microsoft's drivers," Clarke says. "The other reason is that it cannot match the performance of Fibre Channel and, therefore, is never likely to be used to support business-critical applications.

"Rather iSCSI will play a supporting role in the datacentre for attaching to the network back-up and other devices where performance is not critical," she says.

Fareham Borough Council is one of the first local authorities in the UK to adopt cut-price iSCSI networking technology for storing, replicating and backing up data. The council's new San is a combination of Ethernet and IP networking protocols using Gigabit Ethernet data access speeds.

"This package is costing £250,000, including the first year's running costs," says Peter Harper, IT manager at Fareham. "A similar system based on Fibre Channel would have cost at least double that."

Along with the physical extra storage capacity delivered by Sans, NAS and iSCSI, survey respondents confirm they need extra help in managing the systems.

The study finds that almost a third of respondents have bought storage management software, capable of a range of measurements from telling a network manager that extra disc capacity is required on the network to predicting a possible outage. During the next 12 months, a fifth of respondents say they will invest in such management packages.

But despite the availability of this software, three-quarters of respondents admit that they will simply buy more disc capacity to make their storage more efficient.

Yet the survey also finds that two-thirds of respondents say their organisations currently use less than 75% of their storage capacity. In the next 12 months, this degree of under-utilisation by companies is expected to drop to 50% as they improve their storage efficiency.

"Storage capacity often runs at less than 50% especially where direct-attached storage is deployed," Clarke points out. "The implementation of Sans and the pooling of storage has enabled organisations to increase utilisation to about 80%.

"Storage consolidation is increasing utilisation, hence the expected drop in the number of companies with utilisation levels below 75%. This figure will fall further," she adds.

But the under-utilisation of capacity illustrates why the likes of suppliers such as Cisco feel companies need help.

"We are approaching the networked phase of the datacentre to improve cost effectiveness and business agility," says Jonathan Gilad, Cisco datacentre networking product manager. "Companies do not want to continually expand their storage capacity when existing systems are only operating at as little as 20% capacity. Networked and shared datacentres can help increase storage utilisation to about 90%."

Gavin Williams, technology infrastructure director at Avanade, the IT joint venture between Microsoft and consultancy Accenture, says the cheap-fix offered by more disc space is understandable. "Buying extra discs is a solution that is known and familiar to organisations, whereas storage management is a more complicated solution that requires firms to stop, think and invest valuable resources into usage and design decisions."

And though users are spending more on management software to guide them through the storage maze, Alastair McAulay, senior IT infrastructure consultant at PA Consulting, says more fundamental problems have to be addressed when it comes to data storage. "As with any IT infrastructure, storage technology gets refreshed regularly. It is only natural that newer technology such as Sans or iSCSI filter their way into organisations over time.

"Unfortunately, the people who deploy the new data storage technology - the IT department - are not the same people who do data management. Data management is typically a function that is closely related to the business itself," McAulay says.

This creates a situation where the considerable advances in storage technology that have occurred over the past three years are not being exploited, McAulay says. IT departments that have implemented the upgrades are still largely not talking properly to the people who manage and analyse the data, he adds.

The situation is compounded by the fact that new storage technologies have fewer expansion constraints than other technologies, illustrated by the way disc space can be thrown at a storage network to tackle a perceived capacity problem. "To consider an analogy," says McAulay, "it is like having a budget for home improvements and only deciding to expand the attic because decisions on re-arranging rooms seem like too much bother."

So irrespective of the technology used for data storage, nothing will get better until the interface between data management personnel and the IT department improves.

Although the complexity of storage planning is worrying 28% of respondents, almost 90% of companies are confident in their staff's ability to handle the organisation's storage requirements, according to the survey. This confidence is perhaps reflected in the fact that 95% of respondents do not outsource their storage systems.

Clarke says, "A few years ago outsourcing storage was increasing in popularity, but there is now a move among organisations to bring storage back in-house. But with the need to retain more data I would expect to see an increase in outsourcing, but only in archived data such as e-mails."

Paul Trowbridge, board director at industry body, the Storage Network Industry Association, feels the trend is not as clear cut. "What is happening is that medium- to large-sized companies are sourcing professional services and management of their storage while retaining the systems on-site.

"At smaller firms there will be a shift towards outsourcing in the coming years, as these companies may be willing to co-locate more of their IT at a professional services provider," says Trowbridge.

The choices organisations now have to make on storage are much more complicated. The days of simply putting data on tapes and storing them somewhere in the building - or at a remote location for safer back-up - are long gone, even though much of the older data ultimately ends up on tape once it has been taken off a network.

The key aim for any business is not to simply build storage capacity, but to deliver easier data access for those that need it; decide on how old, duplicated and useless data can be removed from the network altogether; and to make sure the IT department is more in touch with the business needs of the organisation when it comes to using and storing data.

The benefits of iSCSI-based storage networks   

By combining SCSI, Ethernet and TCP/IP, an iSCSI solution delivers four key advantages: 

  • Builds on stable and familiar standards since many IT staff are familiar with the component technologies 
  • Creates a storage network with a reduced total cost of ownership since installation and maintenance costs are low 
  • Provides a high degree of interoperability because it reduces disparate networks and cabling by using regular Ethernet switches instead of special Fibre Channel switches which are much more expensive 
  • IP data can travel over the global internet and, therefore, there are no practical distance limitations, unlike established Fibre Channel-based networks for instance.

Who will be the first companies to use iSCSI?   

Organisations with: 

  • Distributed IT 
  • Significant data growth in the past five years 
  • Proliferation of servers in divisional, departmental and workgroup environments 
  • The business requirement to consolidate data storage and management for these environments, to improve operational efficiency, data availability, and storage resource management.

Read more on Networking hardware