Looking ahead: Top 10 storage trends of 2009

analysis
Jan 2, 20094 mins

Solid state, whole-disk encryption, virtual disk management, commodity clusters, and green storage will get real in 2009

1. More flash-based SSD. We’ll see a lot more SSD in use, as major players such as Intel, Seagate, and Samsung start shipping enterprise-oriented SSDs in quantity. While prices will initially limit use to small tier 0 arrays, prices will drop fast enough, and drive lifetimes will improve fast enough, to push adoption into the mainstream before year’s end.

2. High-bandwidth options will hit price points low enough to encourage adoption. 10Gbps iSCSI, 10Gbps Infiniband, and 8Gbps Fibre Channel have all been available for some time, but high prices for HBAs and switch ports have limited adoption. Prices continue to drop, though, and 2009 should see widespread adoption of these standards. Don’t expect to see any one standard predominate. 10Gbps iSCSI will grow in volume, Infiniband will move beyond its current stronghold in HPC into the mainstream, and 8Gbps FC will continue to sell well.

3. Whole-drive encryption will become more widesread, but management will continue to be a headache. With major vendors offering BIOS-level encryption of whole drives for laptops, desktops, and enterprise systems, adoption will grow rapidly, driven by the fear of being sued if user or customer data is lost in another well-publicized debacle. Admins’ well-justified fears of losing data permanently if keys can’t be recovered will be addressed by key management systems, either from storage vendors like Seagate, from specialized vendors like Voltage Security, or from management system vendors like IBM Tivoli, but cross-platform adoption will continue to be problematic.

4. More, better storage for less. Performance will continue to improve, with 6Gbps SAS-2, better PCI-e RAID adapters, and high-speed (8 to 10Gbps) interfaces to storage arrays becoming affordable, while prices will continue to come down rapidly. Thin provisioning, snapshots, support for virtualization, and ‘self-healing storage’ will be available even on inexpensive systems.

5. More deals between virtualization vendors and storage vendors to allow one-console provisioning and management of virtual disks. As interest in virtualization continues to grow, more and more storage vendors will announce agreements with virtualization vendors to integrate management of storage with management of virtual servers. A true cross-platform, enterprise-wide integrated platform that allows for multiple virtualization platforms and multiple storage systems may become available from third-party vendors such as Scalent, but most storage admins will have to live with homogeneous environments to manage virtualization and storage from a single console.

6. Clustered storage will begin to move beyond pilot projects. Storage systems made up of large numbers of commodity Intel boxes with 4 to 8 drives (versus one storage controller with dozens or hundreds of drives) have promised high availability, quick rebuild times in case of node failure, and exceptional performance, but have mainly found a home in academic HPC systems. The promise of inexpensive, high performance storage will encourage more mainstream storage admins to look into this technology.

7. E-discovery and archiving will grow in importance as more legal cases depend on them. The potentially high costs of not being able to provide documentation of procedures being followed (or not) are being made clear by judgments against companies who ignored the need for an e-mail archiving and e-discovery system. That such systems can also enforce HR or legal policies on the use of e-mail and allow for self-service recovery of lost e-mails is just an added bonus.

8. Progress will be made in storage management, but an enterprise-wide, single-pane storage management system will continue to be a dream. SNIA (Storage Networking Industry Alliance) standards such as eXtensible Access Method (XAM) have lots of potential, but adoption will be slow. As per usual, the key vendors are all participating on the various committees, but reluctant to actually cooperate to the degree necessary to enable heterogeneous storage management.

9. The economy and green initiatives will drive power-saving systems. The cost savings gained by using SSDs will not be sufficient to drive adoption in enterprise systems in 2009, but we will see increased use of 2.5-inch high-efficiency drives, and vendors will add features like spin-down or sleep, especially in tier 2 or tier 3 near-line storage. Standards for evaluating actual energy savings will move the battle from marketing brochures to engineering initiatives.

10. There will be a battle between storage silos and SANs with better utilization. Often political boundaries within IT as well as practical, application-oriented limitations have prevented SANs from being efficiently utilized across multiple hardware and software platforms. Having a different SAN connected to each server defeats the purpose of SANs, but sharing storage among various admins with responsibilities for e-mail or databases or file servers on different operating systems can become really problematic. This is one area where the promise of SANs has not been fulfilled. Although some admins may give up and return to direct attached storage, others will be driven by the desire to utilize existing investments (especially when capital budgets are slashed) to find ways to get better utilization out of SAN systems.