(no subject)

This paper proposed a novel approach for point-to-point communication
abstraction to provide services like multicast, anycast, and
mobility. To ease the deployment of such services, the approach is an
overlay-based Internet Indirection Infrastructure that offers a
rendezvous-based communication abstraction (i3).

The purpose of i3 is to provide indirection; that is, it decouples
the act of sending from the act of receiving. Instead of explicitly
sending a packet to a destination, each packet is associated with an
identifier; this identifier is then used by the receiver to obtain
delivery of the packet. It allows the receivers to create, at the
application level, services such as mobility, anycast, and service
composition out of this basic service model. Moreover, the
infrastructure can give responsibility for efficient tree
construction to the end-host.

i3 can be used by many applications to achieve the more general
communication abstractions of mobility, multicast, and anycast. For
mobility, the host is assigned a new address when it moves form on
location to another. The mobility could be address by just updating
each of host's existing triggers. The sending host need not to be
aware of the mobile host's current location or address. It's easy to
achieve multicast in is too. Creating a multicast group is equivalent
to having all members of the group register triggers with the same
identifier id. For anycast, i3 also could address it to be very
effective. We also could use i3 to some other services, like service
composition, heterogeneous multicast, server selection, large scale
multicast.

i3 was designed to be robust, self-organizing, efficient, secure,
scalable, incrementally deployable, and compatible wit legacy
applications. To demonstrate the feasibility of this approach, the
overlay network was built based on the Chord lookup system. In
addition, the authors have developed two other applications:
providing transparent mobility to legacy applications, and a large
scale reliable multicast protocol.

This paper is well-written with clear presentation and good
organization. Its main contribution is to propose a novel
infrastructure. The main point of this infrastructure is to decouple
the sending and receiving which solve the big problems in some
services like multicast, mobility whose problems are not totally
solved by IP-layer solution and application-layer solution. It think
there is a problem is this infrastructure. Routing is achieved by i3
servers, which means i3 servers are responsible for the delivery of
packets. So the performance is determined by the performance of the
servers. Another problem is that if i3 is deployed, how much is the
cost, not only adding the servers in the networks, but also all the
applications should use i3 protocol.
Received on Wed Nov 22 2006 - 15:44:23 EST