How Do You Back Up the Museum of Modern Art?
As visual art evolves on the internet, MoMA is having to adapt the way it conserves both its 630,000 square foot of gallery space and its archives.
Image: Chrissy Hunt/Flickr
New York City's Museum of Modern Art boasts one of the premier art collections in the world. In terms of 20th century art, it's not even a contest. The museum's collection contains Picasso's 1907 pre-cubist masterpiece, "Les Demoiselles d'Avignon." Any day of the week it is possible to go the galleries and see Salvador Dali's iconic "Persistence of Memory," with its melting clocks, or Vincent van Gogh's "Starry Night." Then there are the contemporary galleries, with video monitors, digital animations, and projections from Paul Chan, Ryan Trecartin, and other cultural rebels pushing into new formats.
As visual art evolves on the internet, MoMA is having to adapt the way it conserves both its 630,000 square foot of gallery space and its archives. The museum is turning toward an old-school form of tape data storage in order to back up its collection of both digital-native works and images of physical objects. But as it turns out, storing the data is only the beginning. Museum staff also have to figure out how to manage the future of artworks when the digital platforms they exist on just might disappear within a decade.
MoMA's digital collection is currently about 90 terabytes in size, but the museum expects that to grow to 1.2 petabytes (1.2 million gigabytes) by 2025. That archive will soon be stockpiled on Linear Tape-Open (LTO), a magnetic tape storage system developed in the 1990s.
One of the people responsible for the peculiar challenge of translating so much multimedia material from all eras of culture into digital bytes is Ben Fino-Radin, MoMA's digital repository manager. Fino-Radin describes himself as a media archaeologist; he came from the internet art community, worked as the in-house conservation specialist at the new media blog Rhizome, and has become well known as an expert on how to make digital materials remain accessible, no matter the era they're from or the hardware they exist on.
MoMA's challenge involves more than just scans of paintings. How do you preserve the hacked Nintendo cartridge used in Cory Arcangel's "Super Mario Clouds"? Video: Cory Arcangel
As digital repository manager, Fino-Radin oversees the museum's digital footprint as well as procedural policy for handling data. Whether it's film, painting, or a hacked Nintendo game (thanks Cory Arcangel), all the art has to get translated into the same medium for storage. Images are carefully scanned and analog videos ripped. MoMA also has to make sure artworks are stable—able to be recalled by the viewers of the future.
"Digitization alone is not preservation," Fino-Radin said. "When we digitize things, it's not like way that Internet Archive digitizes books, ripping out the spines and scanning as quickly as possible. It's much more delicate and considered." (August Bourré, who worked as a book scanner for the Internet Archive, tweeted that the Internet Archive used non-destructive scanning practices, and that it was actually Google who would rip the spines from books.) The biggest ongoing project at the museum these days? Digitizing the entirety of Andy Warhol's film archive—over 500 hours of 16 millimeter film. "All in all that'll be just half a petabyte," Fino-Radin said.
Back It Up
After digitization, all those files have to be put somewhere. "The problem with digital preservation is that there's no permanent form of storage, it just doesn't exist," Fino-Radin said. Formats change, companies fail, and data gets corrupted. The best current answer is the resolutely physical LTO Ultrium system, which Fino-Radin is currently transitioning to. The actual tape that holds the encoded data is are stored in cases that look like zip disks, which will then stored in MoMA's basement.
"It will be this giant hulking black box," Fino-Radin said. "Plus a second system at the MoMA Queens art storage facility, and a third copy at the Celeste Bartos Film Preservation Center," MoMA's media vault in Hamlin, Pennsylvania.
The best thing about the LTO system is that as the technology improves, the storage capability of the same amount of physical material increases exponentially in a perfect echo of Moore's law. Unlike, say, a USB drive or memory stick, the entire system doesn't have to be swapped out. "When you do upgrades you have to replace the tape drives that read the tapes but the rest of the infrastructure stays the same," Fino-Radin said. In 2006, LTO held just over 500 GB of storage per cartridge. In 2010 it hit 1500, and in 2012 (named LTO-6), 2500. As updates continue and the digital collection grows, it will still fit on the same tape.
So why not just use the cloud? MoMA's collection just might fit on the largest Dropbox account ever. But there are other dangers to take into account when subscribing to a storage service instead of doing it yourself. "The most basic reason is cost," Fino-Radin said. "When you store things in the cloud, you don't purchase storage, you rent it. With Amazon's cloud, you pay a monthly fee for the capacity you need. When you purchase capacity you own it."
Even with the required migrations, LTO is much cheaper. "We projected that our preservation requirement is that we have three copies in three different geographic locations, which you can do with Amazon Cloud. But it would cost upwards of $10 million more than to store the same data on our own infrastructure," Fino-Radin said.
Then there's the issue of being locked into a particular company that's vulnerable to cyber attacks, internal instability, bankruptcy, or anything else that could take its servers offline. "When you store things in the cloud, you are dependent on the companies you are storing them with to be around in 10, 15, 20 years. Amazon isn't going away, but let's say they did. You have to get all your data out of their system," the digital conservator said. "You always have to have an exit plan."
Amazon has encrypted their data server-side since 2013, but it can't be protected the same way that a bank of tapes in MoMA's own basement can be. The museum's backups are behind a firewall and only accessible via a private network. The files themselves are "write once, read many," Fino-Radin said. "You can write data but can't erase or change data; it's incredibly permanent." This guards against all-important artworks getting stolen, hacked, or altered, so they'll stay the same for future generations—so many digital Mona Lisas.
The Future of Art Storage
MoMA plans a decade out with its digital storage needs, but it's not so concerned with storage infrastructure as what it will be allowed to do with the data it receives. "It really often is driven by rights agreements," Fino-Radin said, who added that in the digital era, the museum has changed its acquisition contracts "to be more permissive with what we can do with the content." That means ripping data, accessing source code, and porting pieces to platforms they might not have originally existed on.
"The really big challenge when writing an agreement is having language that says you can do all this stuff, but not having a date so specific it becomes obsolete," Fino-Radin said. The institution can't specify a particular platform like YouTube since "in 25 years it will probably be irrelevant." The trick is thinking long-term. He offers some basic digital file guidelines that might go as well for a personal music stash as the world's best collection of modern art. "Is it lossless, is it uncompressed, does it have good metadata embedded in it?"
These rules might be what the right glues and solvents are for more traditional conservators working on paintings. They guarantee that future conservators know what they should and shouldn't do to a piece, hewing to an artist's exact intentions. Petra Cortright's early videos, for example, depend on the context of YouTube (she priced them according to how many views they received), so when the platform disappears, that same environment will have to be emulated. Data decays and becomes outdated just as certain pigments do.
In the end, conservators just might end up longing for the comparative ease of maintaining a physical painting in a plain-old white gallery—no server space needed.
Clarification 7/20/15: This post was updated to add a tweet from August Bourré to clarify that the Internet Archive does not use destructive book scanning practices.