Established in 2003, Queensland Brain Institute (QBI) is a leading center for neuroscience research. It is one of the University of Queensland’s flagship research facilities with a growing team of almost 300 research neuroscientists across 33 laboratories. QBI’s research focuses on the cellular and molecular mechanisms that regulate brain function to aid development of new therapeutic treatments for neurological and mental disorders.
QBI’s storage system has to scale extensively to enable the institute to store, protect, and access tens of terabytes of data daily to support cutting-edge research. Its legacy tape technology could not provide ongoing, reliable, and timely access to research data or sufficient data density to support limitless innovation. QBI needed an economical, extremely scalable, fast, and secure alternative to its legacy tape libraries and tape drives within its tiered, storage architecture.
In 2010, QBI upgraded the server, tape library, and metadata storage in its Oracle Optimized Solution for Tiered Storage Infrastructure, replacing smaller components with two larger, faster and more highly-available StorageTek SL3000 modular library systems from Oracle. It also upgraded its legacy servers to Oracle’s Sun Fire X4270 servers for greater reliability, and with Oracle’s Sun Flash Accelerator F20 PCIe Card for exponentially faster daily metadata backups. In 2012, to further enhance its tiered storage architecture, QBI implemented 10 StorageTek T10000C tape drives from Oracle to replace its much lower-capacity legacy Quantum LTO4 tape drives.
QBI can now offer its users highly scalable storage capacity and continuous availability. In addition, it realized an average improvement of 30% more data compression, 66% faster response time, 99% faster metadata backup, and peace of mind regarding data security. The solution also improved QBI’s ability to grow, attract world-leading scientists, and meet stringent funding conditions.
Researchers can retrieve their data 66% faster with StorageTek T10000C tape drive, enhancing workflow and productivity. With legacy tape technology, retrieving a 15 gigabyte confocal microscopy image from LTO4 tape drives took 125 seconds. The same files now load on a user’s desktop in 42 seconds. As an optimized solution, all components are known to work together, further increasing data access.
“StorageTek T10000C retrieves data much, much faster than our legacy system. We’ve witnessed compressed read and write rates of 350 megabytes per second, which significantly accelerates file retrieval and archiving times,” said Jake Carroll, senior information technology manager, QBI.
By implementing Oracle’s Sun Flash Accelerator F20 PCIe Card, QBI can complete metadata backups in just 12 minutes per file system, instead of 22 hours—99% faster—greatly enhancing performance and substantially reducing recovery point objective.
By implementing StorageTek SL3000 and StorageTek T10000C, and utilizing StorageTek SAM to automate movement of data between disk and tape, QBI greatly increased its storage capacity.
“Our storage systems have to support extreme scalability. Our previous tape libraries were not large or fast enough to cope with the tens of terabytes of data that move through the system daily. For example, our brain genomics division generates as much as 72 terabytes of data with each next-generation sequencing run. Without the throughput that the StorageTek T10000C delivers, we would not be able to get data to tape quickly enough, and our file systems would fill up.”
The system also provides advanced compression capabilities. QBI has achieved 2x compression—a 30% improvement over the legacy storage environment. The tape drives in the tiered storage solution enable up to a 12x compression rate on raw text files generated by a genomic sequencer compared to a 4x to 5x compression rate with the previous architecture.
StorageTek SAM’s policy-based multiple copy feature works together with StorageTek T10000C tape drive’s Data Integrity Validation feature and tape protocol verification technology, as well as StorageTek Tape Analytics’ proactive monitoring capability, to provide an extremely safe and reliable archiving environment. Scientists know that their valuable research data is protected against loss or corruption, and QBI can comply with its funding sponsors’ stringent data safety and long-term storage requirements.
“With a vast amount of critical research data on our StorageTek SAM file systems, not only can we can get it back in a timely and verifiable way, but we can get any version of it back for the user. It’s like a giant time machine,” Carroll said.
“The tape protocol validation technology in StorageTek T10000C coupled with Oracle Solaris 11.1 and StorageTek SAM 5.3 allows us to quickly verify the consistency of the media and decide to migrate the data to ensure its integrity. The Data Integrity Validation feature allows us to hold our hand to our hearts and say to our users that, when you put your data into the file system, it is guaranteed to be the same when it comes back out because of the consistency checks taking place in real time, all the time. That level of trust is really important to our scientists, and only Oracle’s StorageTek solutions can deliver it,” Carroll explained.
In a high-output research environment, system downtime can be an expensive disruption to critical research operations. To ensure continuous data access, QBI relies on StorageTek SL3000 modular library system’s dual robotic redundant arms and on StorageTek SAM’s ability to maintain multiple copies of the data and to retrieve the data from any device.
“For example, a genomics run can cost tens of thousands of dollars, so we can’t afford for our file systems to go down while these complex instruments are transmitting data. If a tape library robotic arm fails, another one can take over to keep the system running,” Carroll said.
“To do maintenance on a tape silo, I can take part of the robotics offline while the rest of it still services the research workload. This fully-redundant, hot-swappable component significantly reduces system downtime,” said Carroll.
While cost control and efficiency are imperative for a not-for-profit organization, setting storage quotas for researchers would sabotage innovation. By pushing more data through to tape and paying less per stored terabyte, Oracle’s tiered storage infrastructure enables QBI to guarantee sufficient storage for all user needs at a much lower cost than running all disk-based solutions.
“Let’s say a user decides to sequence a cucumber and generates 100 terabytes of data. It’s very hard to offer them 100 terabytes of disk. SAM software and our new tape technology provides a seamless layer that looks like 100 terabytes of disk but might be 10 terabytes of tape and 90 terabytes of disk. Users are not concerned because they see a uniform interface. It’s all one big, glued-together object to them,” Carroll said.
“Tape media is much, much cheaper than disk-based solutions at this scale, so there’s a huge savings opportunity here. I would also need five-to-six more storage racks in the data center to accommodate the same storage density that I now have, and I’m saving approximately 15,000 watts of power in real time by not having to cool and maintain a large number of disk arrays,” Carroll said.
Further, Oracle’s StorageTek T10000C tape drive is a more cost-effective solution for QBI than current or even next-generation open, linear tape (LTO) technology.
“Currently, I don’t see the value proposition in the LTO standard in my line of work. For almost the same cost as the next technology that the LTO group will release, StorageTek T10000C can give me greater raw data storage capacity per cartridge at higher streaming rates,” Carroll said.
QBI attracts world-leading scientists and continues to produce ground-breaking insights and discoveries. In 2012, QBI opened a new research center to investigate dementia caused by aging, a center that will expand into five independent research groups by 2015. It would be unable to take on such large initiatives without the high capacity and affordability at the scale achieved with the new tape technology and the tools to dynamically create copies and access data from a tiered storage solution.
“You gain a competitive advantage by being able to see more and have your hands on more at the same time. Knowledge is power, and the more stored knowledge, the more power our users have to fulfill their research ambitions,” Carroll said.
QBI chose the Oracle Optimized Solution for Tiered Storage Infrastructure because it has deep confidence in the engineers who created the StorageTek SAM software and in the optimized solution’s tried and tested reliability.
QBI chose StorageTek T10000C over LTO6 tape drives from IBM and HP because it offers better capacity and is better suited to QBI’s high-throughput environment. In addition, StorageTek SAM, together with StorageTek T10000C’s DIV feature, assures the integrity of all data no matter how old. It selected Oracle’s StorageTek SL3000 modular library system over competitor products because of its scalability and high availability. QBI also had great trust in the capabilities and responsiveness of its local Oracle support team after seven to eight years of working together.
“At the moment, Oracle is the only company that can offer archive and hierarchical storage management software, tape library, and drive products with this specific data integrity validation and data assurance technology, and with this level of tape and storage density,” Carroll said.
QBI implemented StorageTek SL3000 and Sun Flash Accelerator F20 PCIe Card in 2010. The organization took delivery of 10 StorageTek T10000C tape drives in November 2012 and had them running in just two weeks.
Oracle Premier Support assisted in the installation of the tape drives, system testing and tuning.
“Oracle’s very experienced StorageTek local site engineers provided invaluable assistance and were highly responsive to our needs at all times,” Carroll said.