On the sleeping nodes, I imagine that there has to be a layer 2 wake
mechanism in there?
Store and forward is easier to consider than proxying/intercepting; in
many cases there's just two separate concatenated HTTP sessions and
hops: sensor to IP gateway, then gateway (often post sensor data
processing) across the terrestrial Internet. In other cases there are
relay nodes in the sensor network adding more hops and more
concatenated HTTP sessions.
Yes, the flexible text nature of HTTP makes it compelling over binary
alternatives. Binary compression of MIME bodies is already carried out
via e.g. gzip content encodings; that doesn't cover the headers. But
there's also something in favour of not compressing headers for debug/
ease of use etc.
FYI, we'll be talking about HTTP (and HTP-DTN) and the various
transports HTTP could be carried over other than TCP in the TSV-AREA
meeting in Stockholm:
Title: Turning HTTP into a full-fledged networking layer
Presenter: Lloyd Wood
Requested time: 10 mins.
On 16 Jun 2009, at 14:07, Brian Frank wrote:
> My take is that HTTP's strength is text, and that we'd want to be
able
> to compress any generic text well, perhaps starting with a common
dictionary
> of HTTP tokens (GET etc.) for the text compressor.
>
> food for thought/discussion in Stockholm?
HTTP-DTN looks really interesting. I especially like the store-
forward model and Content-Source and Content-Destination headers.
One troubling aspect of Chopan is how to handle PUT/POST to a
sleeping nodes. Perhaps a store-forward model might be a better fit
than trying make it a caching model?
The idea of using a common dictionary for compressing any text
(including maybe even the body) might be really interesting. The
beauty of HTTP is that provides a transport for any MIME typed data,
so we can pass binary data, text data, or new versions of compressed
data. For me, the open ended nature of how a resource is
represented on the wire is the most compelling aspect of HTTP.
DTN work: http://info.ee.surrey.ac.uk/Personal/L.Wood/saratoga/
<http://info.surrey.ac.uk/Personal/L.Wood/><[email protected]>
_______________________________________________
6lowpan mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6lowpan
