Neurome Inc.

Biomedical research firm Neurome Inc., which conducts studies of mice brains for pharmaceutical companies, is currently in the process of expanding its datacenter for neuroscience research. Over time, it plans to have one throat to choke when things go wrong, instead of four or five.

Luckily (?) for IBM Corp. (NYSE: IBM), it is the chosen throat. Systems and software from Network Appliance Inc. (Nasdaq: NTAP), Microsoft Corp. (Nasdaq: MSFT), and Compaq (now Hewlett-Packard Co. [NYSE: HPQ]) -- all currently installed in Neurome's data center in San Diego, Calif. -- are being taken out of the production environment, to be used for development projects.

"We are going with a single vendor from a philosophical standpoint, so we get the servers, the backup, the disk systems, and the database software from one company," says Dr. Warren Young, president and CTO of Neurome. "That way there is much less finger-pointing when things go wrong. Right now, NetApp points to Microsoft, who points to Compaq, and it goes round and round. This way, IBM is the company we call and that's it."

Neurome will use the IBM eServer p690 system to process 3D models of mouse brains and IBM TotalStorage FAStT500 storage server to provide up to 7 terabytes of storage capacity to maintain its 3D atlas databases of genetic and proteomic data.

Young is still evaluating IBM's DB2 database software against its installed Oracle Corp. (Nasdaq: ORCL) database and IBM Tivoli's backup software versus its currently installed backup software from Veritas Software Corp. (Nasdaq: VRTS). "Tivoli has some great features, but you need an IBM consultant to come in and install it, there are pros and cons with each," he says. Neurome also had problems running Oracle8i on its NetApp filers -- although that was more of a NAS issue in general (see NetApp on Red Alert and NetApp: 'Thanks, Microsoft!').Overall, Young believes, purchasing everything from a single supplier should solve a lot of their problems.

One of Neurome's biggest hurdles is storing and accessing the vast amounts of data it is gathering from 3D models of mouse brain structures. The amount of data is large because of the precision of the images, which go down to the sub-cellular level. "The finest resolution would expand the size of the mouse brain to five football fields," Young notes.

A single mouse brain apparently requires 10 Gbytes to 1 terabyte of storage -- which is a lot of storage for one mouse. Essentially what Neurome is doing is making models of mice brains that mimic human conditions, in order for drug companies to do tests on them. Neurome received its first multimillion-dollar research contract from Elan Pharmaceuticals.

The company has been in business for about a year and half and currently has over 8 terabytes of data. It expects this to increase by about 4 terabytes a year.

This growth bodes well for IBM, which is making a concerted effort to tap into the biotech market with packaged storage offerings designed to combat the exponential growth of data in this sector (see IBM Targets Drug Makers).At the Drug Discovery Technology 2002 conference in Boston this week, IBM unveiled three bundled storage networking solutions. These are combinations of pre-existing disk drives, tape storage, and servers that have been tested to ensure interoperability and have been designed to gather scientific data into a single repository that can be accessed by disparate members of a research team, IBM says.

For small research groups, IBM has made available a package including the IBM TotalStorage FASt500 mid-range disc for sharing data between servers, the IBM TotalStorage Linear Tape Open (LTO) midrange tape storage device for backup and archiving, and an IBM eServer p630 running IBM's flavor of Unix, AIX.

IBM is actively encouraging research organizations to use its DB2 database software and server management software from Tivoli, but purchasers will not be required to use that software in order to use the storage bundles, it says.

Big Blue estimates the storage opportunity in life sciences at $4.6 billion in 2002, growing to $6.9 billion by 2004. It also says that over 7 petabytes -- or 7,000 terabytes -- of storage are needed each year by 100 universities and labs doing genomic research, which offers a glimpse of the massive data storage requirements facing this emerging industry (see Oxford Cures Storage Ills