What are you looking for ?
Infinidat
Articles_top

UCSC Scientists Develop Solutions for Long-Term Storage

Using building blocks consisting of a HDD, a small low-power processor, a flash memory card, and an Ethernet port

This article was written by Tim Stephens, from the University
of California, Santa Cruz

UCSC Computer Scientists Develop Solutions for Long-Term Storage of Digital Data

pergamum3002


The
team that developed Pergamum includes graduate students Kevin Greenan
and Mark Storer and associate professor of computer science Ethan
Miller.

 

Although the digital age is well under way, one crucial
detail remains to be worked out–how to store vast amounts of digital
information in a way that allows future generations to recover it.

"The problem is how to build a large-scale data storage system to
last 50 to 100 years
," said Ethan Miller, associate professor of
computer science in the Baskin School of Engineering at the University
of California, Santa Cruz.

Tape libraries are widely used for data storage, but digital tape
has many shortcomings as an archival medium. Miller’s group has come up
with a new approach, called Pergamum, which uses hard disk drives to
provide energy-efficient, cost-effective storage. The declining cost of
hard drives has made them more competitive with tape, and they offer
numerous advantages for searching and retreiving data. "It’s like the
difference between a VCR and TiVo
," Miller said.

Pergamum, named after the ancient Greek library that made the
transition from fragile papyrus to more durable parchment, is a
distributed network of intelligent, disk-based storage devices. The
team that developed it includes UCSC graduate students Mark Storer and
Kevin Greenan, along with researcher Kaladhar Voruganti of NetApp
(formerly Network Appliance), a company that focuses on storage and
data management solutions.

Archival storage is a big issue for businesses, partly due to legal
requirements for the preservation of financial and business records,
and also because data mining strategies can turn stored data into a
valuable resource. Long-term storage is also a growing issue for
individuals who are filling their personal computers with digital
photos, movies, and documents.

"There is a risk that an entire generation’s cultural history could
be lost if people aren’t able to retrieve that data
," Storer said.
"Everyone is switching to digital cameras, but we’ve never demonstrated
that digital data can be reliably preserved for a long time
."

Pergamum has attracted a lot of attention from industry since Storer
presented it at a leading conference in the field, the USENIX
Conference on File and Storage Technologies (FAST ’08), held in San
Jose in February. Robin Harris, an industry consultant who writes an
influential blog called StorageMojo, called the Pergamum paper his "favorite FAST ’08 paper".

The researchers designed the system to provide reliable,
energy-efficient data storage using off-the-shelf components. It also
has the ability to evolve over time as storage technologies change.
"You want to avoid ‘forklift upgrades,’ where you have to get rid of
the old system and transfer all your data to a whole new system
,"
Miller said.

According to Storer, businesses are beginning to recognize that
archival storage is very different from simply backing up their data.
"A backup is a safety net–you hope you won’t need it. Archival data
you do want to use–it’s a valuable resource and you want to be able to
mine it for information
," he said.

Tapes work well for backups, in which data are written once, rarely
read, and not kept indefinitely. But archival data should be easy to
read, query, browse, and search, and tape has inherent weaknesses in
these areas. Existing disk-based systems offer excellent performance,
but rely on power-hungry central controllers.

"Energy usage is a big issue, so a lot of our effort in designing
Pergamum focused on dramatically reducing power use
," Miller said.

Pergamum uses individual building blocks consisting of a hard drive;
a small, low-power processor (like the chip in an iPhone); a flash
memory card; and an Ethernet port. These units, called "tomes," are
connected using relatively inexpensive Ethernet switches.

"Each tome is like a minicomputer, but with very low power demands,"
Miller said. "When not in use, it can shut down almost completely."

Even when active, the devices use very little power (less than 13
watts), which can be delivered over the network using Power over
Ethernet technology. As a result, each unit is essentially a
self-contained box with a network connection. The flash memory provides
low-power, persistent storage so that many operations can be performed
without activating the hard drive.

For reliability, Pergamum uses two levels of redundancy–within and
between disks–to protect from both disk failures and errors in writing
data to a disk (so-called ‘latent sector errors’). Tomes can be easily
added to expand the system or to replace failed disks. And if hard disk
drives become obsolete in 10 years, Pergamum won’t suffer the same
fate. The system doesn’t care what the actual storage medium is, as
long as the device can implement the simple protocol that will allow it
to function as part of the network.

"In 50 years, the devices might use holographic storage," Storer
said. "As long as you can wrap the new storage medium in this
intelligent layer that speaks the protocol, it can participate in the
network
."

Pergamum is one of several related projects being developed by
researchers in the Storage Systems Research Center (SSRC) at UCSC’s
Baskin School of Engineering. The center’s other archival storage
projects include Deep Store, which dramatically reduces the amount of
space required to store data, and POTSHARDS, which provides long-term
secure storage using ‘secret splitting’ instead of traditional
encryption. "Both of these projects would be compatible with Pergamum",
Miller said.


Articles_bottom
AIC
ATTO
OPEN-E