|
In the day-to-day operation of a data center, a multitude of tasks must be
performed to ensure that storage systems actually deliver the performance
and reliability that
an enterprise expects. While these tasks may seem conceptually simple --
such as giving higher priorities to more important tasks, archiving
little-used data to less-expensive media, or keeping required records
unmodified -- their sheer volume can often overwhelm
an understaffed center or force the hiring of additional highly trained
storage administrators.
We are studying ways to automate many storage management processes
so administrators can concentrate on those tasks for which their skills
are critically required.
An example of a product that addresses these requirements is IBM's
storage management product, the
IBM TotalStorage Productivity Center. Our team contributed significantly to the release of
this product, and continues to deliver advanced storage management functionality to it;
we are also extending some of the technologies used in storage management into the broader area of systems management.
In collaboration with the Almaden Computer Science function, we are doing
research on a next-generation achive solution that improves data accessability
while optimizing data center costs.
We are also focused on working with customers and IBM services in order to
better understand and serve our customers' needs.
Storage Management:
Our vision is to develop integrated storage network analytics & performance management tools to
aid policy-based autonomic resource provisioning of storage networks on IBM TotalStorage
Productivity Center (TPC), an open solution designed to manage complex, heterogeneous
storage infrastructure.
Intelligent Data Storage:
Organizations are under increasing pressure to properly
retain and manage their records.
The Intelligent Data Storage (IDS) project is a collaborative effort
between the storage systems and computer science teams to develop
techniques and solutions for handling new and emerging forms of
archive.
Mirage:
Mirage is an intelligent out-of-band storage optimization engine that provides autonomic analytics for optimal storage resource allocation, migration, consolidation and tiering in complex virtualized data center environments. An important philosophy of Mirage is to provide time-tested solutions that minimize configuration changes into the distant future.
SMART:
Storage Management Analytics and Reasoning Technology (SMART) is a self-evolving, corrective action engine
that optimizes storage-resource allocation in a fully automated, cost-efficient way so most clients experience
predictable performance in their accesses to a shared, large-scale storage utility.
Storage Management for Petascale Supercomputers:
IBM’s Programmable Easy-to-use Reliable Computing System (PERCS), chosen by the
Defense Advanced Research Projects Agency
(DARPA),
is one of two system designs to be developed and demonstrated as part of phase III of the
High Productivity Computing Systems program
(HPCS).
Windsor W. Hsu, Soumyadeb Mitra and Marianne Winslett.
Trustworthy Keyword Search for Regulatory-Compliant Records Retention.,
VLDB'06 - Very Large Data Bases Conference, Sep 2006. (Best Paper Award)
Kristal Pollack and Sandeep M. Uttamchandani.
Genesis: A Scalable Self-Evolving Root-Cause Analysis Framework for Storage Systems.,
ICDCS'06 - The 26th International Conference on Distributed Computing Systems, March 2006.
Li Yin, Sandeep M. Uttamchandani, Madhukar R. Korupolu, Kaladhar Voruganti and Randy Katz.
SMART: An Integrated Multi-Action Advisor for Storage Systems.,
USENIX'06 - USENIX Annual Technical Symposium, March 2006.
Sandeep M. Uttamchandani, Li Yin, Guillermo Alvarez, John D. Palmer and Gul 'Agha.
CHAMELEON: a self-evolving, fully-adaptive resource arbitrator for storage systems,
USENIX'05 - USENIX Annual Technical Symposium, April 2005.
Qingbo Zhu, Windsor Hsu.
Fossilized Index: The Linchpin of Trustworthy Non-Alterable Electronic Records,
SIGMOD 2005 - International Conference on Management of Data, 2005.
Windsor Hsu, Alan Jay Smith, Honesty Young.
The automatic improvement of locality in Storage Systems,
ACM Transactions on Computer Systems, 23(4), 2005.
Aameek Singh, Madhukar Korupolu, Kaladhar Voruganti.
Zodiac: Efficient Impact Analysis for Storage Area Networks,
FAST'05 - 4th USENIX Conference on File and Storage Technologies, 2005.
Sandeep Uttamchandani, Kaladhar Voruganti, Sudarshan Srinivasan, John Palmer, David Pease.
Polus: Growing Storage QoS Management Beyond a 4-year old Kid,
FAST'04 - 3rd USENIX Conference on File and Storage Technologies, 2004.
Michael Kaczmarski, Tricia Jiang, David Pease.
Beyond backup toward storage management,
IBM Systems Journal 42(2), 2003.
Selected publications from all three research areas of
Almaden Storage Systems can be found here.
Advanced Storage Systems
File Systems
More information on IBM-wide storage systems research can be found at
IBM Research - Storage Systems
|
