Upgraded Data Archives Position Lab for Exascale Era
Lawrence Livermore National Laboratory (LLNL) is now home to the world’s largest Spectra TFinity system, following a complete replacement of the tape library hardware that supports Livermore’s data archives. Housed behind Sierra—the world’s 2nd fastest supercomputer—the new tape library helps the Laboratory meet some of the most complex data archiving demands in the world and offers the speed, agility, and capacity required to take LLNL into the exascale era.
LLNL’s high performance computing (HPC) users produce massive quantities of data, averaging 30 terabytes (the equivalent of 58 years’ worth of digital music) every day. Not only must the data be kept indefinitely, much of it is classified and must be protected with the highest standards of security. The solution is to archive—something Livermore Computing (LC) has been doing since 1967. But, as with all technology, data storage hardware becomes antiquated and expensive to maintain, thus requiring significant upgrades and maintenance at fairly regular intervals (every several years). A well-run archive is a constant maelstrom on the backend, including data migration to newer technology and smooth, uninterrupted responsiveness to user needs.
“Moving to a completely new tape library is an enormous undertaking that’s only done once every decade or two,” says Todd Heer, who oversaw the procurement process. “This particular procurement was even more notable because we chose to go with a new vendor—Spectra.
The new Spectra TFinity systems are novel in two ways: First, the tape cartridges they contain are stored differently, in terapacks (that is, 10 cartridges per pack), which means that more data can be stored in a smaller footprint.
“This physically denser storage solution is much better for LC,” says Heer. “The Spectra libraries take up significantly less floor space than our previous libraries, which allows us to be more efficient and agile as the big computers come and go.”
Second, the new tape library comes in a “rack form factor” (similar to traditional HPC systems, which hold their various components in multiple rack enclosures). This feature gives Livermore the option of growing the system, whenever necessary, simply by adding a frame or two. The current main configuration is 23 racks, which contain 128 TS1155 (“enterprise” class) tape drives and 19,575 slots. (There are an additional six racks nearby that store secondary copies of data on “midrange” class tape.) It is capable of storing 294 petabytes of uncompressed data, which is enough space to hold the entire written works of humanity, in every language, since the beginning of history six times over. The new technology represents a 50% or better density improvement over outgoing LC tape media.
Heer explains, “Our plan is to procure maintenance in five-year increments and add just enough racks to allow for near-future growth, which realizes efficiencies in cost and overall size. We’ll also be able to capitalize on ever-increasing tape-drive capacities. In time, the library capacity will multiply simply as a function of generational drive technology improvements.”
One aspect that has not changed is LC’s reliance on tape drives rather than a disk-based option. Contrary to popular belief, half-inch-tape storage technology continues to advance and thrive. According to Heer, it is the most economically priced storage medium, with the lowest bit error rate, and best power profile compared to other currently available storage technologies. While not ideal for RAM, or memory, quick random accesses, the trick is to leverage its strengths by applying the technology in write-once, read-seldom environments.
Another unique feature of the TFinity tape library is its custom front panel “skins,” or coverings. Working with a graphic designer at Spectra, LC staff designed the panels to depict the evolution of LLNL supercomputers, research, and storage systems over the last six decades—offering visitors and tour groups a unique visual history of Livermore’s HPC ecosystem.
Over the next year, LC’s Data Storage Group will be planning and executing the wholesale swap of media and data from the old tape libraries to the new—a process that occurs, remarkably, with no downtime to the users.
“We’re really proud of this new system and the important role it plays for our HPC users today and in the future,” says Heer.
The TFinity tape library is funded by the Advanced Simulation and Computing (ASC) program and serves the tri-lab community.