"Mark Gritter Department of Computer Science Stanford University
"Tuesday, November 29th, 8:30-9:30 AM Gates 498 Refreshments served at 8:15 AM.
"Abstract:
"The Internet's continued growth has demonstrated the need for a secure, scalable, reliable naming service. Today's Domain Name Service is viewed as a second-class service, subordinate to packet delivery, but naming is a necessary component of nearly every Internet transaction. Even content distribution networks which aim to provide high availability and performance rest upon a naming infrastructure which is vulnerable to long delays, failures, and attacks.
"Name-Based Routing performs name lookup hop by hop using the same set of routers and network connections needed to connect to the resource named. Instead of centralized name servers, Name-Based Routing relies upon "content routers" which exchange naming information in a manner similar to that in which BGP exchanges address routing information.
"This thesis outlines the design and behavior of name-based routing. It demonstrates that Name-Based Routing achieves near-optimal latency for connection setup when accessing content and services over the wide area. It shows that NBR eliminates off-path resources, improving reliability by reducing network dependencies to the minimum necessary for a given level of redundancy. It demonstrates how NBR may be scalably implemented and deployed given the expected growth of the Internet. Finally, the work demonstrates that NBR is resistant to large-scale denial of service attacks."
Anything which improves upon the mass of cruft and bogosity that is DNS is a good thing, in my book. As well as getting rid of the need for centralized name servers.
I have a strange feeling VeriSign may not feel the same way, though...
Woo hoo! Stanford no less! Congratulations to him, you and lh user="timprov">! I was support staff for albionwood during both of his master's degrees. It's a household project.
Very true. Although somebody still has to be in charge of avoiding name clashes. (There are ways to get around that, too, it's just not part of my thesis.)
(And I did, in fact, just ring the bells. -The Bells of Dublin- by the Chieftains is on the 5-CD player at work, and it starts with church bells, so I switched over to it and cranked the volume. Woot!)
Because, having been the thesis-writer in this scenario, I know exactly how difficult one of our breed is to live with, and how deeply relieved and joyful y'all are.
no subject
Date: 2005-11-29 07:27 pm (UTC)no subject
Date: 2005-11-29 07:27 pm (UTC)no subject
Date: 2005-11-29 07:34 pm (UTC)no subject
Date: 2005-11-29 07:35 pm (UTC)no subject
Date: 2005-11-29 07:37 pm (UTC)What was the thesis?
no subject
Date: 2005-11-29 07:38 pm (UTC)no subject
Date: 2005-11-29 07:40 pm (UTC)no subject
Date: 2005-11-29 07:41 pm (UTC)no subject
Date: 2005-11-29 07:43 pm (UTC)no subject
Date: 2005-11-29 07:45 pm (UTC)no subject
Date: 2005-11-29 07:45 pm (UTC)So, do we have to call him Herr Doktor now?
no subject
Date: 2005-11-29 07:49 pm (UTC)"Mark Gritter
Department of Computer Science
Stanford University
"Tuesday, November 29th, 8:30-9:30 AM
Gates 498
Refreshments served at 8:15 AM.
"Abstract:
"The Internet's continued growth has demonstrated the need for a
secure, scalable, reliable naming service. Today's Domain Name
Service is viewed as a second-class service, subordinate to
packet delivery, but naming is a necessary component of nearly
every Internet transaction. Even content distribution networks
which aim to provide high availability and performance rest upon
a naming infrastructure which is vulnerable to long delays,
failures, and attacks.
"Name-Based Routing performs name lookup hop by hop using the same
set of routers and network connections needed to connect to the
resource named. Instead of centralized name servers, Name-Based
Routing relies upon "content routers" which exchange naming
information in a manner similar to that in which BGP exchanges
address routing information.
"This thesis outlines the design and behavior of name-based routing.
It demonstrates that Name-Based Routing achieves near-optimal latency
for connection setup when accessing content and services over the
wide area. It shows that NBR eliminates off-path resources,
improving reliability by reducing network dependencies to the minimum
necessary for a given level of redundancy. It demonstrates how NBR
may be scalably implemented and deployed given the expected growth
of the Internet. Finally, the work demonstrates that NBR is resistant
to large-scale denial of service attacks."
no subject
Date: 2005-11-29 07:50 pm (UTC)no subject
Date: 2005-11-29 07:52 pm (UTC)no subject
Date: 2005-11-29 07:54 pm (UTC)no subject
Date: 2005-11-29 07:56 pm (UTC)no subject
Date: 2005-11-29 08:01 pm (UTC)no subject
Date: 2005-11-29 08:02 pm (UTC)no subject
Date: 2005-11-29 08:08 pm (UTC)no subject
Date: 2005-11-29 08:14 pm (UTC)P.
no subject
Date: 2005-11-29 08:15 pm (UTC)no subject
Date: 2005-11-29 08:18 pm (UTC)no subject
Date: 2005-11-29 08:21 pm (UTC)no subject
Date: 2005-11-29 08:21 pm (UTC)Anything which improves upon the mass of cruft and bogosity that is DNS is a good thing, in my book. As well as getting rid of the need for centralized name servers.
I have a strange feeling VeriSign may not feel the same way, though...
no subject
Date: 2005-11-29 08:24 pm (UTC)no subject
Date: 2005-11-29 08:30 pm (UTC)no subject
Date: 2005-11-29 08:34 pm (UTC)no subject
Date: 2005-11-29 08:37 pm (UTC)Quite a timely topic, too.
no subject
Date: 2005-11-29 08:48 pm (UTC)no subject
Date: 2005-11-29 08:59 pm (UTC)(And I did, in fact, just ring the bells. -The Bells of Dublin- by the Chieftains is on the 5-CD player at work, and it starts with church bells, so I switched over to it and cranked the volume. Woot!)
no subject
Date: 2005-11-29 09:28 pm (UTC)no subject
Date: 2005-11-29 09:31 pm (UTC)no subject
Date: 2005-11-29 10:00 pm (UTC)no subject
Date: 2005-11-29 11:31 pm (UTC)no subject
Date: 2005-11-29 11:40 pm (UTC)no subject
Date: 2005-11-29 11:47 pm (UTC)no subject
Date: 2005-11-30 01:30 am (UTC)no subject
Date: 2005-11-30 04:26 am (UTC)no subject
Date: 2005-11-30 04:46 am (UTC)Um. That probably did not deserve an exclamation point.
no subject
Date: 2005-11-30 04:48 am (UTC)no subject
Date: 2005-11-30 11:08 am (UTC)no subject
Date: 2005-11-30 11:08 am (UTC)no subject
Date: 2005-11-30 11:09 am (UTC)Veryvery pleased that this is over, in other words.
no subject
Date: 2005-11-30 08:12 pm (UTC)*hugs* all around and a *ear scritch* to the puppy