[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
RE: Proposed charter, please comment
- To: Markus Hofmann <hofmann@bell-labs.com>, "Maciocco, Christian" <christian.maciocco@intel.com>
- Subject: RE: Proposed charter, please comment
- From: "Tomlinson, Gary" <gary.tomlinson@cacheflow.com>
- Date: Fri, 26 Jan 2001 15:01:35 -0800
- Cc: "'Kobus van der Merwe'" <kobus@research.att.com>, tme@21rst-century.com, John Martin <jmartin@netapp.com>, Abbie Barbir <abbieb@nortelnetworks.com>, Mark Day <markday@cisco.com>, cdn <cdn@ops.ietf.org>
- Delivery-date: Fri, 26 Jan 2001 15:03:00 -0800
- Envelope-to: cdn-data@psg.com
On Friday, January 26, 2001 @2:30 PM PST Markus Hofmann wrote:
>Mechanisms such as re-direction etc. should NOT rely on multicast
>capabilities of the network. These mechanisms must be able to operate
>over unicast networks, given that multicast is rarely deployed these
>days. Of course, such mechanisms might take advantage of multicast
>where available, but then it should be optional.
>
>The same way, multicast should not be REQUIRED for streaming. It can
>be used for better efficiency, but a streaming solution should not be
>designed based on the assumption that (network-level) multicast is
>available everywhere.
I agree completely with Markus on this one. In fact, the distribution and
subsequenet delivery of streaming by surrogates for live events is a form of
application level multicast in its own right that isn't dependent upon the
network layer. Network multicast support can make this more efficient, but
its not needed for the system to work.
My .02
Gary