At $78.9 billion, the federal government's proposed IT budget for fiscal year 2013 is 0.7% less than the current budget. That's four years in a row that the federal IT budget has been flat, but there's been no letup in the growth of data or the need to store it. During this same period, the feds' data storage requirements have been growing 30% to 40% a year, gobbling up scarce IT funding.
We hear a lot about the imperative to reduce costs and make the federal government more efficient--data center consolidation and "cloud first" are two prominent examples--but less has been said about a key technology: storage.
But the move to cloud computing and the growing importance of big data and information sharing are creating a sense of urgency around data storage. Networked storage systems such as network-attached storage and storage area networks are part of the answer, but they're not enough.
Most agencies have adopted a tiered storage architecture that involves different technology components for different functions and data sensitivity. This architecture includes various physical media (NAS, SAN, and others), as well as policies and services that govern functions within the storage environment.
Some data typically requires real-time or near-real-time access, which can be accommodated by the architecture. With NAS, data files are stored and accessed using standard file systems and protocols, including Network File System for Linux and Unix clients and Common Internet File System for Windows clients. Agencies may choose to spend more on SANs, where data files are stored on highly available devices. Magnetic hard disk drives that store data are typically architected in a way to include highly available RAID. Solid-state disk drives, which are extremely costly, are reserved for data that requires very fast access. For budget-strapped agencies, solid state is largely out of reach.
For slower data access, at significantly lower costs, magnetic tape drives in robotic libraries can handle archived data or data that's protected as part of a backup strategy. They're also used to archive files no longer needed in operational environments but that must be retained in accordance with the federal electronic records retention policy.
Your archiving requirements should govern the movement of data files from fast-access storage devices to slow-access devices. This will help reduce the growth of the faster, but more expensive, storage systems. As needed, data files moved to a tape archive can always be copied back to fast-access data storage using a data recovery service.
A backup service copies data to an indexed storage location on slow-access devices such as magnetic tape drives. These copies are stored for data protection purposes. If the files are needed because of accidental deletion or corruption, they can be recovered.
For agencies that need continuous data protection, a replication service copies all changes to data and files from one storage subsystem to a secondary subsystem. This service can provide a zero recovery point objective to ensure that agencies never miss a beat. A recovery service copies files from the archive or the backup storage location to the operational system with fast-access storage. Recovery also enables data logged during replication to be retrieved from the logs and copied to the operational system.
All of these services require storage media and data governance policies. With the booming growth of data across government, anything less than a well-conceived plan will drain an agency's IT budget.