[Dev] Existing p2p systems for getting offline node back up to date?

Mike C. Fletcher mcfletch at rogers.com
Wed Nov 13 20:20:47 PST 2002


I may be mistaken, but don't distributed hashtables try to keep the 
messages pressed into them for as long as possible w/out regard to their 
delivery status? That is, rather than being able to retire an item when 
it reaches it's recipient, the engine is trying to store a document for 
as long as possible, but normally can't give any guarantee that anyone 
has accessed it before it is dropped?

In other words, if I commit a message to a distributed hash-table, it 
might be available when the user I'm targetting comes online (except in 
the common case that they don't check their mail for 2 weeks), but the 
systems don't (out of the box) support tracking whether the object was 
accessed by the correct user (or do they), how about dropping the object 
when such an event occurs (to save message-storage-space for messages 
not-yet delivered), and what about notifying the sender in the event of 
failure?  That is, one of the critical requirements for a messaging API 
is that it provide some sort of guarantee to either succeed, or fail 
with notification (and thereby allow recovery).

Not saying it can't be done (I think it would be fairly doable, heck, a 
few approaches popped into my head as I wrote this sentence), just that 
as-of-yet, I haven't come across an off-the-shelf (open-source) solution 
that provides everything you need to implement an end-to-end messaging 
system with the requisite guarantees.

Anyway, I started off down this little embarrassment by saying that the 
technical details appear to have been solved around and about various 
projects.  I got deeper by saying there doesn't appear to be a 
ready-made OS solution. I will now state, categorically, and for the 
record, that I don't really care any more ;) , well, not at this point 
in time, anyway.  The original point was "don't worry about the tech at 
this point, define what we actually want to have happen", and I think 
I'll have to retreat there and let the subject be.

Mike

Paul Snively wrote:

>
> On Wednesday, November 13, 2002, at 12:08  PM, Mike C. Fletcher wrote:
>
>> I'm getting convinced (I've been researching this for a while now) 
>> that the functionality _doesn't_ yet exist as Open-Source in the form 
>> I thought (stand-alone P2P messaging server mechanism with message 
>> queuing and re-routing).
>>
> Not true at all. Practically any "distributed hashtable" does 
> precisely what we've been discussing. See 
> <http://oceanstore.cs.berkeley.edu>, 
> <http://www.csse.monash.edu.au/~pfh/circle> (this one's in Python!), 
> <http://research.microsoft.com/~antr/Pastry>, 
> <http://sourceforge.net/projects/khashmir> (also in Python).. I could 
> go on, but that's a good start. :-)
>
>> Enjoy,
>> Mike
>
>
> Best regards,
> Paul
>
> _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
>
> Open Source Applications Foundation "Dev" mailing list
> http://lists.osafoundation.org/mailman/listinfo/dev
>

-- 
_______________________________________
  Mike C. Fletcher
  Designer, VR Plumber, Coder
  http://members.rogers.com/mcfletch/






More information about the Dev mailing list