Re: [vwrap] Status and future of the VWRAP working group

[Dzonatas Sol <[e-mail]>]

In review of HTTPBIS (w/SPDY), TLS, and XMPP for forward-thinking 
cross-development with VWRAP, the one main concern I have with those and 
where I think VWRAP needs to focus is the language-agnostic approach in 
connections and transmissions. If we could move VWRAP more to the core 
and define default type system for basic data, then we can use the 
streams from the others to continue progress. The current complication 
to just add the others is how there is often allowances for JavaScript 
to control the connection or affect it in any way, which to me is 
similar to hardware-hacking (and easily breaks stream-ability, 
complicates batch queries, expects direct browser awareness, and more 
issues could be said we've known from real hardware-hacking).

Those JavaScript allowances obviously have had the web-browser viewpoint 
paradigm in place, and VWRAP has had the virtual world paradigm. The 
problem with virtual world paradigm is that from one VW to another VW is 
like reinventing-the-Internet. I think VWRAP needs to refocus the 
"world" paradigm to useful patterns to virtualize and "wrap" these other 
IETF & RFC standards within XML. I don't mean redo every scheme, yet 
there are obvious patterns in the other WGs (& RFCs) that can use 
capabilities as deployed and present in VWRAP (yet to be fully outlined) 
rather than reinvent protocols. VWRAP can use the streams, presence, and 
TLS possibilities of the other WGs (& RFCs).

As I have pointed out here: 
http://groups.google.com/group/spdy-dev/msg/2fe946257e157684
I continue to make the case/proposal for pivotal data. Imagine all 
queries being sent through a proxy (like with SPDY enabled) that then 
bundles (i.e. "wrap"s) up the data into a stream such as this basic flow:

<? barebones XML std built-in IETF DTD ?>
<XML><IETF wg="vwrap">
<TLS>AVATAR:furryA
<HTTP>GET ...</HTTP>
<HTTP>PUT ...</HTTP>
<HTTP>DELETE ...</HTTP>
</TLS>
<TLS>AVATAR:furryB
<HTTP>POST ...</HTTP>
</TLS>
</IETF><IETF wg="...">
...
<IETF></XML>

If you think of the above flow as batched for the proxy to perform on 
the agent server, then the http queries could be used to manipulate 
inventory for furryA and furryB. The agent server could host WebSockets 
that viewers connect to such that would have those actions processed and 
then further bundled into the above for the region server to preform. 
Those http queries can actually be locally done on the region server, 
which then avoids transmission of the queries. The responses can be 
further bundled in return to the agent server.
This basically projects the queries that would normally take place on 
the viewer sides into batches on the region.

It only looks complicated until you consider the proxy does the http 
queries instead of the viewer or browser. Follow this pattern to 
"virtualize" monolithic browsers and viewers. We can design local proxy 
that acts as the (local) agent service for the viewer/browser. As IPv6 
is fully in operation in June, we can use IPv4 clouds to be these 
proxies/agent-servers to the IPv6 world. The VWRAP enabled clouds adds 
purpose, not only to proxy, yet to stream, frame, and bundle queries 
between the IPv4 "region" and the IPv6 "region".

Some already consider "the cloud" as the avatar/asset teleport device in 
regards to virtual worlds.

Beside power consumption, one way I visualize these virtualization 
techniques is imagine your two robots/avatars on Mars and here on Earth 
you have a 3D surrounded room full of partitioned X-Windows and sound. 
There is a satellite as the proxy/agent-server. How can we use current 
protocols optimally to display here what is seen there? On top of that, 
add security for asset servers.

Of note, the charter shows June 2011 for "Digital Assets". We could turn 
some of those HTTP requests, in the example above, directly into DAEs. 
When the proxy receives any DAE, the asset is stored into the cache. 
When the browser/viewer requests that asset item, the local cache 
already has the item and the browser/viewer possibly does not need to do 
further queries to the Internet. Essentially, detangled the web.

On 04/30/2011 09:45 AM, Barry Leiba wrote:
>> That said, we need to be leading this discussion on consensus that can
>> be documented and posted. �And we need to focus on that and accomplish
>> it soon, for a vague but near-term value of "soon".
>>      
> We had a good bit of discussion in early April.  Now that we're at the
> end of April, and the discussions seem to have stopped for the last
> couple of weeks, I'd like a progress report.  Has there been any work
> on coming to consensus on the direction the group wants to take?  Any
> progress on consensus for the contents of an intro/overview document?
>
> Barry, as chair
> _______________________________________________
> vwrap mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/vwrap
>
>    


-- 
--- https://twitter.com/Dzonatas_Sol ---
Web Development, Software Engineering, Virtual Reality, Consultant