By Salvatore Salamone
January 15, 2003 | In 2003, life science companies will be able to acquire greater amounts of storage capacity for less money and take advantage of new high-performance networking technology to access stored data.
The need to organize, find, and efficiently store ever-increasing amounts of data is a constant challenge for anyone in the life sciences. “We have eight mass spectrometer machines that produce 60GB of data per hour per machine running around the clock,” says Lloyd Segal, president and CEO of Caprion Pharmaceuticals Inc.
About a week’s worth of Caprion data would seem unmanageable, but it’s nothing out of the ordinary in a life science company. Anecdotally, many industry experts expect data within life science companies to double every 6 to 12 months.
Fortunately, acquiring the raw capacity to store the data will not be a problem, the same as in previous years. In December, an IDC report on third quarter sales of storage systems found vendor revenues were down 3 percent compared to the same quarter last year. However, even with a drop in revenue, about 10 percent more capacity in raw terabytes of storage was shipped. Not surprisingly, IDC expects this trend to continue.
For many companies simply adding capacity is not sufficient. Adequate infrastructure must also be in place to access such large volumes of data. “It’s not just the size of storage that matters, it’s the size of the fire hose,” says Segal. “Having a billion-gallon tank that’s accessed through a straw is no good.”
To reduce the potential bottleneck between applications server and the stored data, some storage vendors are adding support for storage area networks (SANs). For instance, late last year Network Appliance Inc., which for years downplayed the usefulness of SANs in favor of Network Attached Storage (NAS), added support for SANs to its product line. SANs divide storage into relatively few logical units, much like a single disk, so users access the disk directly. NAS organizes storage into logical sets of files, which are accessed using file sharing and protocols such as Common Internet File System.
InfiniBand Products to Arrive
This year, in addition to increased support for SANs, products that support the relatively new InfiniBand standard will begin to make their way to market. InfiniBand is a higher performing replacement for the commonly used Peripheral Component Interconnect (PCI) technology to connect most servers and storage devices today. PCI offers 8 Gigabits of throughput per second versus 20 to 60 Gb/for InfiniBand.
Analysts are predicting a strong market for InfiniBand over the next few years. The Yankee Group expects the market for InfiniBand-enabled servers and storage devices will grow from about zero last year to $1.5 billion by 2006.
More than 70 companies plan to bring InfiniBand products to market this year, according to the InfiniBand Trade Association. Some of the companies, such as Hewlett-Packard, EMC, Network Appliance, Sun Microsystems, and IBM, are household names in the storage arena. But many other are not. For instance, companies such as DivergeNet, OmegaBand, and Paceline Systems Inc. will roll out InfiniBand solutions as well.
InfiniBand is expected to play a major role in moving data between stand-alone storage devices and individual server clusters For example, Sandia National Laboratories is using a switch from Paceline Systems in a pilot network to assess InfiniBand as interconnection technology for its high-performance computing clusters.
Beyond faster storage connections, life science companies will also see a trend toward smarter storage applications. “Advanced functions, such as volume management and storage virtualization, can be implemented in the fabric,” says Dan Tanner, an analyst at the consultancy the Aberdeen Group. “Storage network buyers will soon find themselves evaluating storage applications and then considering which networks run them.”
One area where vendors are adding more intelligence is with regulatory compliance. For example, storage vendors are adding management software to meet the confidentiality and access control requirements of the Health Insurance Portability and Accountability Act (HIPAA) and 21 CFR Part 11.
In December, IBM announced new storage networking systems designed specifically for sharing, managing, and securing clinical trial patient information such as electrocardiograms and other digital images.
Many storage vendors, such as IBM, EMC, and DataMirror, are adding features to comply with 21 CFR Part 11 by providing audit trails and access control to the data.