Re: [p2pi] ALTO and caching (Was: Re: Charter and problem statement)

[Laird Popkin <laird at pando.com>]

There are different entities in different p2p networks that might do cache discovery.

The current caching vendors provide:
- In-line 'cache discovery' where the cache is in-line in traffic, caching data and responding to requests rather than passing the requests for cached data on to other peers. This is fragile and doesn't scale well.
- DPI-based 'cache discovery', where a DPI box inspects the p2p traffic and tells the p2p cache server about p2p tranfers, which the cache server can join into. This is also fragile, involving protocol reverse engineering.
- Client-based cache discovery, where each p2p client on each download start would perform (theoretically) one cache check for each cache vendor, and use whatever comes back. This is complex and inefficient.
- Server-based cache discovery, where each p2p network knows where p2p cache servers are and tells appropriate peers to connect to the caches. This is the most efficient approach, but requires ISPs to inform p2p networks of cache locations and rules (e.g. IP prefixes) for assigning users to caches, or equivalent lookups.

- Laird Popkin, CTO, Pando Networks
  mobile: 646/465-0570

----- Original Message -----
From: "Vijay K. Gurbani" <vkg at alcatel-lucent.com>
To: "Stanislav Shalunov" <shalunov at shlang.com>
Cc: p2pi at ietf.org
Sent: Wednesday, July 16, 2008 4:08:56 PM (GMT-0500) America/New_York
Subject: Re: [p2pi] ALTO and caching (Was: Re: Charter and problem statement)

Stanislav Shalunov wrote:
> On Jul 16, 2008, at 8:13 AM, Vijay K. Gurbani wrote:
>> Who do you see doing cache discovery?  The ALTO server or the
>> individual peers?
> 
> Peers mainly.  The ALTO server might use these same means as a 
> convenience, but one can think of it as being configured with this 
> information.

OK; good.  So we have peers that mainly perform cache discovery
using some protocol (more on this below.)

>> I believe that if cache discovery is moved to the realm of
>> the ALTO server, it becomes more of cache "dissemination" problem
>> than of cache "discovery."  In other words, the ALTO server
>> will be pre-configured by the ISP  -- or if it is not operated by
>> the ISP, will use other means -- to maintain a set of known caches.
>> It has been suggested that caches "register" to the ALTO server.
>> I think that this is probably not a good idea for a variety of
>> reasons.
> 
> I don't see why we need to standardize how ALTO servers are configured 
> in this regard.  

You misunderstood.  I was not saying that we standardize how ALTO
servers are configured (we should not go down that path, obviously.)
The point of my paragraph above was to determine if the ALTO server
is the only entity that does cache discovery, do we need to
specify a protocol at all?

So, it looks like the following statements apply (for sake of
sanity, I will take a leap of faith and assume that we have an
actual ALTO server instead of representing it a URL of a BGP dump):

Discovering caches and other resources:

   1) Peers may discover caches using a protocol X. The protocol
    X may be as simple as using well known DNS techniques to
    bootstrap the discovery process (i.e., something akin to
    DNS-SD, BEP-22, etc.)

   2) The ALTO server may be asked by the peers to discover caches
    using certain extensions to the ALTO protocol Y (Note: Y != X).
    When so asked, the ALTO server can use protocol X to discover
    those caches on behalf of the peers, and using additional
    information provided by the peer.

   3) The ALTO server may be asked by applications to discover
    other resources using certain extensions to the ALTO protocol Y.
    When so asked, the ALTO server can use protocol X to discover
    such resources on behalf of the applications, and using additional
    information provided by the application.

Yes?

Thanks,

- vijay
-- 
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
2701 Lucent Lane, Rm. 9F-546, Lisle, Illinois 60532 (USA)
Email: vkg at {alcatel-lucent.com,bell-labs.com,acm.org}
WWW:   http://www.alcatel-lucent.com/bell-labs
_______________________________________________
p2pi mailing list
p2pi at ietf.org
https://www.ietf.org/mailman/listinfo/p2pi
_______________________________________________
p2pi mailing list
p2pi at ietf.org
https://www.ietf.org/mailman/listinfo/p2pi