S. Raman and S. McCanne. Generalized data naming and scalable state announcements for reliable multicast. Technical Report CSD-97-951, University of California, Berkeley, June 1997. 119  Home/Search   Document Not in Database   Summary   Related Articles   Check

....that message s content, and thus, whether it is important or not. Further research has proposed the parallel use of two multicast protocols: An unreliable protocol used for payload and a reliable protocol used to convey meta data describing the content of data messages sent on the payload channel [12]. Using this information the receiver may evaluate the relevance of lost messages. Our approach is inspired on this principle, but exploits the semantic knowledge at the sender side instead. 3 Semantic reliability The basic idea behind our approach is that in a distributed application some ....

S. Raman and S. McCanne. Generalized data naming and scalable state announcements for reliable multicast. Technical Report CSD-97-951, University of California, Berkeley, June 1997.

....in hops, or distance in delay. 3.3. 2 Data Naming Data naming protocols are an application level approach to organizing data efficiently for end to end transport, enhancing the expressibility of an application s requirements to a lower level network service, such as a reliable transport protocol [46, 13]. Data naming helps large scale applications to organize their data flows efficiently, allowing efficient dissemination to interested hosts. As applications increase in size, directing data to interested users requires some organization at the application layer. An open issue is the design of the ....

....network layer; in other words, deciding what the network layer can offer data named applications in terms of special routing support. This issue is a natural intersection of work on data naming and AIM. Data naming protocols can be used to separate data into separate data flows (i.e. containers [46, 13]) and positional routing, anycasting, and streams can be used to support on the fly and lightweight data flows within single multicast group, providing greater application control over data dissemination. 70 3.3.3 Reliable Multicast Architecture Reliable multicast protocols provide end to end ....

[Article contains additional citation context not shown here]

S. Raman and S. R. McCanne. Generalized data naming and scalable state announcements for reliable multicast. Technical Report UCB/CSD-97-951, UC Berekely, June 1997.

....This is followed by a discussion of how we used the upcall mechanism to increase interactivity and the selective reliability to reduce bandwidth requirements. 4.3. 1 Data Naming and Ordering To support application defined reliability, the SRM framework uses a generic, structured naming scheme [35]. The naming structure divides the data naming space into a tree with containers at the inner nodes and individual Application Data Units (ADU s) at the leaf nodes. To discover metainformation like the set of all containers, the framework uses a scalable session announcement MediaBoard: A Shared ....

RAMAN, S., AND MCCANNE, S. Generalized Data Naming and Scalable State Announcements for Reliable Multicast. Tech. rep., University of California, Berkeley, CA, June 1997. CSD-97-951.

....to the application about messages that have been lost. For instance, it has been proposed the parallel use of two multicast protocols: An unreliable protocol used for payload and a reliable protocol used to convey meta data describing the content of data messages sent on the payload channel [16]. Using this information, the receiver may evaluate the relevance of lost messages and explicitly request retransmission when needed. Our approach is inspired on this principle, but exploits the semantic knowledge at the sender side instead. As we will explain later in the text, this allows us to ....

S. Raman and S. McCanne. Generalized data naming and scalable state announcements for reliable multicast. Technical Report CSD-97-951, University of California, Berkeley, June 1997.

....in application data units (ADU) which then become the sole unit of processing, transmission, and control across all network levels. ALF applications cannot rely on a simple, monotonicallyincreasing sequence space to distinguish interleaved ADUs. They require an alternate method of naming data [20], 21] While data naming was introduced as a necessary component of ALF based applications, we take a more general view of naming: any large scale application requires a good organization of its application entities and content for network communication when supported 3 by addressing or ....

S. Raman and S. R. McCanne, "Generalized data naming and scalable state announcements for reliable multicast," Tech. Rep. UCB/CSD-97-951, UC Berekely, June 1997.

....in application data units (ADU) which then become the sole unit of processing, transmission, and control across all network levels. ALF applications cannot rely on a simple, monotonicallyincreasing sequence space to distinguish interleaved ADUs. They require an alternate method of naming data [24], 25] While data naming was introduced as a necessary component of ALF based applications, we take a more general view of naming: any large scale application requires a good organization of its application entities and content for network communication when supported by addressing or filtering. ....

S. Raman and S. R. McCanne, "Generalized data naming and scalable state announcements for reliable multicast," Tech. Rep. UCB/CSD-97951, UC Berekely, June 1997.

....data units, and decisions for making repair requests and replies entirely to the application. Although ALF is a useful design principle, the current implementation of SRM results in developers having to reimplement a lot of the functionality in each application. Researchers in the MASH project [R97] are currently looking into a generalized hierarchical ADU naming scheme, which might possibly alleviate some of these problems by allowing the SRM toolkit to deal with issues such as request reply suppression and relieve the applications of that burden. 5.2 Multicast address allocation In ....

Suchitra Raman, "Generalized Data Naming and Scalable State Announcements for Reliable Multicast" Class project, UC Berkeley 1997

....in hops, or distance in delay. 3. 2 Data Naming Data naming protocols are an application level approach to organizing data efficiently for end to end transport, enhancing the expressibility of an application s requirements to a lower level network service, such as a reliable transport protocol [25, 26]. Data naming helps large scale applications to organize their data flows efficiently, allowing efficient dissemination to interested hosts. As applications increase in size, directing data to interested users requires some organization at the application layer. An open issue is the design of the ....

....network layer; in other words, deciding what the network layer can offer data named applications in terms of special routing support. This issue is a natural intersection of work on data naming and AIM. Data naming protocols can be used to separate data into separate data flows (i.e. containers [25, 26]) and positional routing, anycasting, and streams can be used to support on the fly and lightweight data flows within single multicast group, providing greater application control over data dissemination. 3.3 Reliable Multicast Architecture Reliable multicast protocols provide end to end ....

[Article contains additional citation context not shown here]

S. Raman and S. R. McCanne, "Generalized data naming and scalable state announcements for reliable multicast," Tech. Rep. UCB/CSD-97-951, UC Berekely, June 1997.

....API design because its flexibility offers applications the opportunity to actively participate in the loss recovery procedure. To avoid ACK implosion, SRM uses NACKs. Receivers detect losses from discontinuities in sequence numbers (or by other means with a generic data naming scheme [112]) and transmit NACKs as a request for retransmission of the lost data 1 . A randomized algorithm determines when a receiver transmits a NACK. These NACKs are multicast to the entire group so that any receiver, in particular the closest receiver with the requested data, may generate a repair in ....

Suchitra Raman and Steven McCanne. Generalized Data Naming and Scalable State Announcements for Reliable Multicast. Technical report, University of California, Berkeley, CA, June 1997.

....the gateway multicasts. Since the full announcement is multicasted, other session directories will receive it and will not need to issue a redundant request when a different client selects the session. This scheme could be realized using the flexible reliable multicast framework described by Raman[20]. 3.2 Proxy Announcements When a client wishes to advertise a session, it locates an appropriate instance of the agent and contacts it via TCP. In order to do so, it must find the unicast address of the agent process. One solution to this problem is for the user to configure the session directory ....

Raman, S., and McCanne, S. Generalized data naming and scalable state announcements for reliable multicast. Tech. Rep. UCB/CSD-97951, University of California Berkeley, 1997.

No context found.

S. Raman and S. McCanne. Generalized data naming and scalable state announcements for reliable multicast. Technical Report CSD-97-951, University of California, Berkeley, June 1997. 119

No context found.

S. Raman and S. McCanne. Generalized data naming and scalable state announcements for reliable multicast. Technical Report CSD-97-951, University of California, Berkeley, June 1997.

No context found.

S. Raman and S. McCanne. Generalized data naming and scalable state announcements for reliable multicast. Technical Report CSD-97-951, University of California, Berkeley, June 1997.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC