When you think of Texas, what comes to mind? 'Big' may be the word. So, you can bet it was a big undertaking for Mike Luter, CTO of Cancer Therapy & Research Center (CTRC) in San Antonio, Texas, to talk to his board of directors about his IP storage strategy.
"The CTRC has 137 members on the board of directors," said Luter. "Many have been on the board for a long time and are former politicians and doctors who are from the pre-computer era."
With digital imaging growing and business continuance growing more critical, Luter had to make a change in the way he managed his storage. As a non-profit, he didn't have oodles of cash to do it with.
The CTRC services about 200 patients daily at four locations, using eight radiation vaults. The locations feed into two data centers -- all utilizing centralized and local servers with no dedicated Fibre Channel. Luter has two Clariion arrays using a Cisco SN5428 iSCSI router over a Gigabit Ethernet MAN. Most of the hospital's research computers use a Mac OS while the larger patient treatment side (about 65% of all computers) run PC/Intel.
Many of the patients that go through the CTRC get MRIs, CT scans for bony tissue and soft tissue. While each of these images takes up a good deal of space individually, the real load comes when the images are fused together to create a 3-D image. The 3-D images give doctors an excellent look into a patient – but they also take up a tremendous amount of capacity. Performance was also an issue. Doctors had to routinely wait about 45 seconds for patient records to open.
"We use about 100 MB per patient per four-week stay. We have about 10,000 patients a year," said Luter. "Business continuance is also critical to this operation. Our model requires a patient treatment every 10 minutes."
Now, 10 minutes may not seem like a lot of time, but in a radiation treatment scenario Luter says you can never get the 10 minutes back. If a patient misses a treatment it can be very detrimental to their health. So, system availability is of the utmost importance.
So, about nine to10 months before the iSCSI standard was ratified, Luter and his team decided to give IP storage a chance to reduce infrastructure costs.
IP could not only help Luter save the $2,000 dollars for each HBA card he needed for 45 servers, but he could also use it for most of the critical applications the hospital needed.
"Performance of our voice over IP, Internet, financial systems and Web-based applications was not affected by iSCSI," said Luter. "We also used cross-gig disk mirroring."
In the past, Luter would always have to boot his SAN right from the HBA. Using iSCSI he now has the capability of remote booting. This enabled him to recover servers in about 10 minutes. Another benefit iSCSI offered was eased OS management.
Taking a chance on implementing iSCSI also paid off in the radiation suites. Doctors who once had to wait 45 seconds for patient data were now treated to the same information in roughly 15 seconds.
With iSCSI Luter has certainly been able to keep up with the 'big' mantra in the "Lone Star" state. Big cost savings, minus big management headaches equals big performance upgrades and a big win for iSCSI all around.
Back to the Denver Health IP storage case study.
For more information on IP storage take a look here: