When it came to having a technology backbone, the IT infrastructure of the much renowned Rajiv Gandhi Cancer Institute and Research Center (RGCI) was rather primitive. Then, in 2009, the medical establishment realized the need for a major revamp, because the extremely slow systems were affecting patient care and productivity. So, not only did the institute change its hospital information system (HIS) but also the entire IT infrastructure—under a project called Transcend (Transformation of service effectiveness and delivery).
The picture problem
In the year 2010, RGCI deployed EMC storage (Dell was marketing EMC then). They purchased storage with about 20TB space then. However, by 2011 Dwivedi realized that they were almost out of space. “It was then that we started looking for a long term solution for picture archival.” The center needed a storage management system that could scale to meet the capacity needed for data intensive images and records.
As a cancer hospital, it was impossible for RGCI to foresee or draw a pattern to patients’ visits. “A patient may visit us the next day, after a few weeks, and maybe after 10 years,” says Dwivedi.
This means the data containing patient medical records and images needs to remain highly available rather than becoming less accessible within archival solutions such as tape libraries.“The biggest challenge in our industry is that we cannot relegate images to secondary storage as they need to be readily available, through convenient access.”
The implementation of the Fujifilm Synapse PACS medical imaging system (that enable RGCI staff to store, view and share images electronically) has provided many benefits including instant access to images to track patient history, and elimination of deterioration in image quality that can occur with film. However, PACS is a data-intensive system and the center’s current storage area network (SAN) was reaching its full capacity. “We wanted to deploy an integrated platform that would support HIS, electronic medical records, and PACS. During planning, we realized that our current storage solution didn’t have the scalability or the price-to-performance ratio that we needed based on our predicted data demands,” says Dwivedi.
With an existing relationship with Dell for the supply of almost all server room components, RGCI evaluated storage management systems from EMC, HP, and Dell Compellent against 40 parameters. Dwivedi says, “The evaluating committee selected Dell Compellent as the most suitable storage solution for our requirements. It delivered on its competitive price-to-performance ratio promise, and the highly scalable architecture meets the expanding data needs of our medical systems,” says Dwivedi. “Our major reasons for going with Compellent were that it is open ended and can easily scale up to 2 petabytes,” says Dwivedi.
The solution
The institute deployed a Dell Compellent storage solution and Dell PowerEdge servers to deliver a stable, scalable platform for the hospital information system, picture archival and communication system (PACS) and electronic medical records. “After a thorough consideration of our needs, we bought a 50TB of storage from Dell and implemented the new storage architecture in November 2012,” elaborates Dwivedi.
RGCI deployed a Dell Compellent storage array and Dell PowerEdge R720 servers powered by Intel Xeon processor E5-2600 product family to run PACS on an Oracle database, with VistA (Veteran Affairs Information System and Technology Architecture) being used for electronic health records.
As a result of this implementation, RGCI can store medical images and patient records electronically, which can be instantly compared to previous images. These images can be quickly shared between consultants and remain highly available during the course of the diagnosis and treatment. “Having medical images and patient records accessible online, reassures our patients because their surgeon is aware of all their history before making an informed decision. With our Dell Compellent storage management system, we have eliminated the potential for image deterioration and patient records can be quickly accessed to ensure treatment begins immediately,” says Dwivedi.
He also reveals that with this solution and its scalability options, RGCI is prepared to meet upto seven years of projected data growth. “We can meet our data volume forecast for the next seven years by simply adding more controllers and disk space without incurring license fees or the cost of expensive components required by other solutions,” says Dwivedi.
RGCI did not face any major challenges during the implementation because they contracted Dell to do everything. “Our point of contact was solely Dell,” explains Dwivedi. Though data migration did require a lot of planning, it did not prove to be very challenging for RGCI. The implementation cycle was about a week to 10 days. “The biggest advantage with Dell is that we get excellent service levels,” adds Dwivedi.
Virtual future
RGCI has already started with server virtualization (on VMware) and they are looking at virtualizing storage too, in the near future. Though they are effectively managing storage through the Enterprise Manager, Dwivedi believes that virtualization will bring more transparency into their storage infrastructure.
Where previously storage was provisioned up front based on estimated size volume, now disk space is only used when data is written. Dwivedi elaborates, “We no longer pre-allocate space which could be used for another urgent need because Dynamic Capacity thin provisioning software delivers a realistic picture of available space rather than provisioned disk space.”
According to Dwivedi, RGCI did not see many tangible differences in the performance. “We were early movers and thus were still maintaining good speeds and access timings on the old storage when we decided to migrate,” he explains. Also, as of now, RGCI is only using about 20TB out of their 50 TB storage, so the response time is fairly good.
Another thing high on their agenda is another instance of Compellent, for easy replication and a cost effective disaster recovery. “We do have a disaster recovery site with CtrlS. However, replicating from Delhi to Hyderabad is proving to be very cost-intensive for us. We realized that if we do our backup from Compellent to Compellent, it would not only be simple, but also very cost-effective. We are also planning to shift from manual back-up to an online module in the next one year,” sums up Dwivedi.