Re: [p2pi] Charter and problem statement

[Laird Popkin <laird at pando.com>]

While not required, the ALTO server could benefit from a consistent identifier of a swarm (and thus piece of content) in a number of ways.
- Different treatment of different content (e.g. optimize some traffic, and not other traffic, or apply caches to some content and not other content).
- Build up a 'state' of a swarm so that the ALTO server can respond more rapidly than if it has to compute each response for each request.

I'll also point out that we shouldn't assume that peers (or trackers) send lists of IP's. They could (and in large swarms, should) send lists of network locations (e.g. IP prefixes, ASNs) with peer information (e.g. number of seeds and leeches) and receive back a weight set of network locations, which the p2p software would apply to specific client IP addresses. This allows for optimizations such as (for example) not needing to make an ALTO request for additional peers in the same network location, since they would all receive the same guidance.

I also think that there's a level of detail about peer state that is managed by the p2p network and need not be sent to and processed by the ALTO server. For example, Pando collects extensive profiles of peers (e.g. firewalled) and uses that to optimize which peers are connected to each other. This profiling and associated logic is constantly evolving, so it does not make sense to try to standardize and and move it into ALTO servers.


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

----- Original Message -----
From: "Stas Khirman" <stas at khirman.com>
To: "Laird Popkin" <laird at pando.com>
Cc: p2pi at ietf.org, "Richard Woundy" <Richard_Woundy at cable.comcast.com>
Sent: Saturday, July 19, 2008 7:42:29 PM (GMT-0500) America/New_York
Subject: RE: [p2pi] Charter and problem statement

> -----Original Message-----
> From: Laird Popkin [mailto:laird at pando.com]
> Sent: Saturday, July 19, 2008 6:04 AM
> To: Stas Khirman
> Cc: p2pi at ietf.org; Richard Woundy
> Subject: Re: [p2pi] Charter and problem statement
> 
> Note that content optimization can also (potentially) be content-specific.
> For example, in P4P, the guidance to the p2p networks can be 'tuned' to
> reflect the specific distribution of content. So for both peer selection
> optimization and for cache location there are use cases that could utilize
> some form of content identifier.
> 
[Stas Khirman] 
Sorry, probably I missed your point - how content information can be used in
peer selection (traffic optimization) ? 

Let's consider a use case when ALTO has no knowledge about additional peers
(such as cache) to "enrich" list supplied by client for optimization. 

ALTO receives list of peers ( and their characteristics) from client and
should send back ordered ( and probably filtered) list of the same peers. 

What kind of information to be used in this process?

IMHO, ALTO needs to know peers characteristics (leeacher, seeder,
firewalled, etc),network topology, utilization, some traffic business rules,
but I fail to see how content information can be useful here at all....




_______________________________________________
p2pi mailing list
p2pi at ietf.org
https://www.ietf.org/mailman/listinfo/p2pi