Re: [Notifications] Reliability

>

When discussing reliability, it would be helpful to be more specific  
as to what kinds of impairments are being discussed. For example, is  
this about the loss of a message due to packet loss or server failure  
or a service model that silently drops notifications if there is no  
receiver available (as opposed to a store-and-deliver-later model)?

I think we have a pretty good handle on the former, with both SIP and  
XMPP offering reasonable models that survive most short-term network  
failures or at least provide a failure indication if delivery is not  
possible.

Henning

> Do you think the reliability requirements for mail-related events  
> (not delivery, not access,  but events) are really significantly  
> greater than the reliability requirements for SIP and instant  
> messaging?   If I was in IM conversation with a colleague and one  
> of their messages to me got dropped, that would seriously derail  
> the conversation and I would not find that acceptable.  I've never  
> noticed that happen, however.
>
> One of the points I was making, which I think you agree with, is  
> that if we're going to build on top of the deployed pub/sub  
> infrastructure, the reliability of that system is what it is.  We  
> can only make certain choices in our schemas to improve on that or  
> break it, without changing the system functionality itself, which I  
> believe we don't want to get into.  In the long run, if people want  
> greater reliability than what we get by default, people will build  
> new systems or strengthen the reliability of their existing systems.
>
> thanks,
> Lisa
>
>
> _______________________________________________
> Notifications mailing list
> Notifications@ietf.org
> https://www1.ietf.org/mailman/listinfo/notifications

_______________________________________________
Notifications mailing list
Notifications@ietf.org
https://www1.ietf.org/mailman/listinfo/notifications