The past few years have been an exciting time for IP Storage, as the technology evolved from a selected number of pilot projects to deployment in production environments across many different markets. The main drivers were all irrefutable customer needs: cost-effective storage consolidation, data protection, and disaster recovery for distributed IT environments. The differentiator that made IP storage a success story was the demand to have these strategies leverage the existing investments in Ethernet infrastructure and expertise.
What is IP Storage?
IP Storage consists of three different protocols:-
- iSCSI (Internet SCSI)
- iFCP (Internet Fibre Channel Protocol)
- FCIP (Fibre Channel over IP)
These allow organisations to use Ethernet to carry their storage data: iSCSI is a means of transporting SCSI packets over TCP/IP to enable Ethernet-based Storage Area Network (SAN); iFCP and FCIP are bridging technologies which provide interconnection of Fibre Channel SANs and TCP/IP networks.
In this article we will focus on iSCSI as this technology is used in 95 percent of IP Storage deployments.
The main benefits offered by iSCSI are the low cost, flexibility, distance, and familiarity associated with Ethernet networking technology. To understand where customers are deploying IP Storage solutions, it is best to start with a general view of the typical enterprise IT infrastructure.
Today the penetration of networked storage as opposed to direct-attached storage (DAS) in an organisation’s core data centre can be as high as 70 percent – the vast majority being in the form of Fibre Channel SAN environments. Core data centres in large organisations are most likely to have both Fibre Channel and Ethernet knowledge in-house. This core data centre may be replicated at a remote facility for disaster recovery purposes, but the more likely scenario is that the disaster recovery facility is one of the company’s regional or departmental data centres.
Regional and departmental data centres usually look quite different from the primary data centre. In the former servers are likely to be smaller and much more numerous. The applications running in these environments may not be mission-critical but they are certainly business critical, and they are most likely to be driving significant data growth, the same data growth that causes a constant staffing and asset management issue. The need for networked storage solutions is likely to be most acute here, although their penetration is only likely to be around 30 percent, breaking down to a fairly even mix of Network Attached Storage (NAS) and small SAN environments. This type of data centre is unlikely to have either storage-focused or Fibre Channel-knowledgeable dedicated staff.
Outside of the regional and departmental data centres there can be a significant number of remote offices. Interestingly, these offices are today likely to employ not only desktop systems, but also a number of small servers – often running small enterprise applications. IT support for remote offices is often an issue and data availability can be a challenge; most companies are struggling with the integration of these offices into their corporate data protection and management environment. Interestingly, smaller organisations are likely to have a similar environment, but the core data centre will look much more like the regional/departmental data centre of a larger organisation – i.e., low networked storage penetration, low Fibre Channel SAN penetration.
IP Storage – Where it Fits
IP Storage solutions greatly broaden the options available to IT executives to address the cost, availability, performance and manageability issues caused by constant data growth, and they accelerate the transition from a direct-attached to a networked storage model.
iSCSI provides the basis for networked storage solutions in environments where Fibre Channel would not be cost-effective, for example in departmental data centres where we find many lower-end servers typically running midrange enterprise applications; this is an excellent price/performance fit for iSCSI over Gigabit Ethernet. In this environment iSCSI is also a popular solution for disk-based storage for regulated or archive data. In such cases disk storage is often based on ATA drives which provide fast data restore with a cost per megabyte comparable to that of small tape silos.
Low end iSCSI solutions are also becoming popular inside SANs in remote offices, as they provide the performance and availability benefits of centralised storage while also giving the option to support file storage; this is where SAN/NAS convergence appears.
Which applications are best suited to IP SAN solutions?
Today, iSCSI is being deployed primarily in environments dominated by mid-range and low-end servers where the applications tend to be mostly business-critical at the departmental level and sometimes mission-critical in smaller enterprises.
Organisations relying on home-grown applications running on smaller servers with internal or direct-attached storage are also finding iSCSI a great storage consolidation solution, since the migration to an IP SAN is affordable and relatively painless.
IP storage is rapidly becoming a mainstream networked storage technology; thanks to IP-native storage SANs can today be deployed in a much broader range of environments and users can enjoy a wider range of price points than ever before. Storage consolidation, simplified data protection, affordable disaster recovery, and improved data management are all major reasons behind IP Storage implementations.
The pace at which IP Storage has gone from an emerging standard to a mainstream networked storage technology is truly staggering and the rate of adoption is quickening. The time is now right for all IT executives to take advantage of the latest networked storage opportunities and find out how IP Storage can address some of the most critical data management pain points in their environments.
By Aad Dekkers, Chair, Storage Networking Industry Association Europe IP Storage Initiative, www.snia-europe.org