Abstract:

The remarkable drop in storage costs makes it possible and attractive
to capture past application states and store them for a long time.
This opens the possibility that kinds of analysis like forecasting,
formerly dependent on data warehouses and temporal databases, can
become available to everyday applications in off-the-shelf data
stores. The challenge is how to organize past states so that they are
``not in the way'' and ``always there'' when needed.

Our  approach, called Retro, integrates a low-level consistent
snapshot system into a data store storage manager, allowing to run
unmodified data store programs against the snapshots, side by side
with programs running against the current state. The approach is
attractive for several reasons. An application can take snapshots
efficiently with any frequency, keep them indefinitely, or
garbage-collect them at low cost, a useful feature in long-lived
systems. A principled methodology derives the snapshot protocols from
the native data store storage manager mechanisms, allowing to
implement the snapshot system in a modular way, without extensive
modifications to the data store internals, making the approach
suitable in off-the-shelf data stores. The talk will describe the new
techniques that underly Retro and present preliminary performance
results from a prototype we built in Berkeley DB, indicating Retro is
efficient, imposing moderate performance penalty on the native data
store, on expected common workloads.


Short bio for Liuba Shrira

Liuba Shrira is a Professor in the Computer Science Department at
Brandeis University, and is affiliated with the Computer Science and
Artificial Intelligence Laboratory at MIT. She received her PhD from
Technion (Israel) what feels like yesterday, working with Nissim
Francez, Michael Rodeh and Oded Goldreich on theoretical aspects of
distributed algorithms. From 1986 to 1997 she was a researcher in the
MIT Programming Methodology Group with Barbara Liskov. She joined
Brandeis in 1997. In 2004-2005 she was a visiting researcher at
Microsoft Research, Cambridge, UK. In 2010-2011 she will be a visiting
researcher at Microsoft Research Asia and a visiting Professor in the
Computer Science Department, Technion. 

Her research interests span aspects of design and implementation of
distributed systems and especially storage systems. This includes
fault-tolerance, availability and performance issues. Her recent focus
is on long-lived transactional storage, time travel (in storage),
software upgrades, byzantine faults, and support for collaborative
access to long-lived objects. 

Liuba Shrira is a member of ACM and of the IEEE Computer Society. She
has been recognized as a Distinguished Scientist by ACM for
"significant accomplishments in, and impact on, the computing field". 



Refreshments served from 14:15 on,
 	Lecture starts at 14:30