[Pl-seminar] Talk @ Harvard: Thursday, February 13, 2003

Mitchell Wand wand at ccs.neu.edu
Mon Feb 3 16:48:35 EST 2003


------- start of forwarded message (RFC 934 encapsulation) -------
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Return-Path: <owner-colloquium at deas.harvard.edu>
Delivered-To: wand at ccs.neu.edu
Precedence: bulk
Reply-To: Carol Harlow <harlow at deas.harvard.edu>
From: Carol Harlow <harlow at deas.harvard.edu>
Sender: owner-colloquium at deas.harvard.edu
To: colloquium at deas.harvard.edu
Subject: Talk - Thursday, February 13, 2003
Date: Wed, 29 Jan 2003 10:45:07 -0500

Harvard University Computer Science Colloquium Series33 Oxford St.,
Cambridge, MA 02138
tel: (617) 496-1440  fax: (617) 495-9837

Colloquium

The Challenges of Content Addressed Storage

Michael Kilian
Centera Chief Technologist
EMC
http://www.emc.com/centera


Thursday, February 13, 2003
 4:00PM
 Maxwell Dworkin G125
 (Ice Cream at 3:30PM - Maxwell Dworkin 2nd Floor Lounge Area)

Abstract
This talk will discuss Centera, a new content addressed storage (CAS)
device for unchanging information.  Unlike conventional storage devices
(e.g., block or file system-based), CAS completely segregates the naming
of an object from its storage location. More to the point, the content
being stored actually defines the address of the object.

This talk will focus on the concepts behind CAS, and particularly
Centera. Because of features like a globally consistent address space
for objects, CAS offers a number of challenges.

a) How globally unique is a content address?
b) How can we implement the complete virtualization of storage CAS
offers?
c) How do we model reliability in a system that heals itself?
d) What are the programming model challenges?

CAS also offers many intriguing possibilities for future architectures.
Given the dramatic decrease in the cost of storage and the ease of
managing truly enormous (and distributed) address space with CAS, should
we bother throwing anything away?  Should we now focus on temporal views
of storage?  What happens when ubiquitous (and infinite) storage meets
ubiquitous (and infinite) bandwidth? Finally, how do we access storage
that is more self-aware of its own content?

By the end of this talk I hope you will not only be intrigued with a new
way of thinking about storage, but also challenged to think of new ways
of applying a new paradigm.

Host:    Professor Margo Seltzer


- --
Carol Harlow
Harvard University
Maxwell Dworkin 343
33 Oxford Street
Cambridge, MA  02138
617) 496-1440
Fax (617) 496-6404


------- end -------


More information about the pl-seminar mailing list