2011/6/30 Pieter Hintjens p...@imatix.com:
Artur, an interesting paper. If you have running code you might want
to refer to that in the spec.
-Pieter
Agree on this. It should probably be part of the guide with some
introduction explaining how it compares to message routing protocols
On Tue, Jun 28, 2011 at 12:06 PM, Martin Sustrik sust...@250bpm.com wrote:
On 06/28/2011 05:57 AM, Artur Brugeman wrote:
It's based on XREP sockets now. I've given quite a lot thinking to it,
and can share my experience.
Please, do so.
Martin
As the letter grew, I've decided to post
i don't understand what this sort of specification is meant to accomplish.
it seems to be a description of a processing flow for a rudimentary job
scheduling system.
but its not complete enough for anyone to use nor implement.
its hard to discern what is 0mq related, and what is job scheduler
Hi, Andrew.
i don't understand what this sort of specification is meant to accomplish.
I said I didn't find a better place to post this on zmq website, so posted
it at rfc.zeromq.org. It is simply a too big for an email message. It is
meant to accomplish availability for people to read it.
it
On Thu, Jun 30, 2011 at 4:27 AM, Andrew Hume and...@research.att.com wrote:
and is the style to talk about such things on the list or on the web page?
List is better, we didn't enable comments on the spec pages because
such comments tend to become outdated very rapidly.
Artur, an interesting
On 06/28/2011 05:57 AM, Artur Brugeman wrote:
It's based on XREP sockets now. I've given quite a lot thinking to it,
and can share my experience.
Please, do so.
Martin
___
zeromq-dev mailing list
zeromq-dev@lists.zeromq.org
Andrew,
On June 24, 2011 08:41:13 PM Andrew Hume wrote:
henry,
i'm not sure i followed all your examples, but i detected a general meme
which others have complained of in the past. i admit to being sensitised
to this issue because i have worked in this area for over a decade now.
just to expand a little on what pieter said, take heartbeats.
many people, when they first confront those, they think the problem is easy.
at some level it is, but often people realise that a networking heartbeat does
not imply application health, and so they end up doing an application level
On 06/28/2011 04:50 PM, Henry Baragar wrote:
Does queue management fall into the realm of messaging or job
scheduling? If it falls into the realm of messaging, then shouldn't the
queue be end-to-end? If it falls into the realm of job scheduling, then
should 0mq have any queues at all?
I
MS I would say the usage of term queue originates from the olden days
MS when 0MQ was still thought to be some kind of equivalent to enterprise
MS queueing products. The queues are more like network buffers today.
Buff Sockets? SoS (Sockets on Steroids)? SSS (Simply Smarter Sockets)?
:)
-- Gregg
Naive approach would be to allow for explicit acks so that there could
be at most one request sent to a specific peer at any single moment.
That would at least allow load-balancer to distinguish busy peers from
idle peers and base the load-balancing on that.
However, that doesn't seem to
it is not synchronous versus asynchronous.
it is somehow deeper than that, but confused because in one
small part of the space, they overlap.
i will take a stab out some text, post it here for comments, then pieter
can think about if, or how, he might add it to the guide.
andrew
On Jun
On Sun, Jun 26, 2011 at 1:06 AM, Andrew Hume and...@research.att.com wrote:
i will take a stab out some text, post it here for comments, then pieter
can think about if, or how, he might add it to the guide.
I'm more than happy to write up any useful patterns in the Guide.
Given enough time,
On Fri, Jun 24, 2011 at 11:11 AM, Henry Baragar
henry.bara...@instantiated.ca wrote:
the slow joiner syndrome...
It has to be handled on a pattern-by-pattern basis. My belief today is
that if you want synchronization over push-pull, you must switch to a
router-based pattern, in the same way as
On 06/26/2011 04:26 PM, Pieter Hintjens wrote:
It has to be handled on a pattern-by-pattern basis. My belief today is
that if you want synchronization over push-pull, you must switch to a
router-based pattern, in the same way as we use routers to synchronize
request-reply workers.
My feeling
On 06/25/2011 02:41 AM, Andrew Hume wrote:
i think this meme is quite common; nearly all instances of complaints
about 0mq's
scheduling and buffering, slow joiners and the like, are examples of
this -- we want 0mq
to trivially do job scheduling for us as well as all the other stuff.
That's
Hi Henry,
I have followed the 0MQ mailing list for about a year, experimented with
0MQ and contributed to the 0MQ adaptor for plack. I like many of the
features of 0MQ, including asynchronous I/O, multi-language support,
fan-out/fan-in connections and end-point connection syntax. But there
Yeah I think Henry gone all out on this one as well. These issues mainly are the
problem of the application developer not 0MQ (or that's how I see it), basically
what 0MQ offers is the infrastructure required to do lots of abstracted
communication tasks in the form and shape of patterns.
I've
On 06/24/2011 07:57 PM, Amr Ali wrote:
Yeah I think Henry gone all out on this one as well. These issues mainly are
the
problem of the application developer not 0MQ (or that's how I see it),
basically
what 0MQ offers is the infrastructure required to do lots of abstracted
communication
On 06/24/2011 09:20 PM, Henry Baragar wrote:
Yes, I have a couple of workload distibution applications with highly
asymmertric workloads (see my message from from 2011-04-04 Re: LRU
broker queue in intuitive way on 3.0
http://lists.zeromq.org/pipermail/zeromq-dev/2011-April/010352.html).
I
On June 24, 2011 03:45:20 PM Martin Sustrik wrote:
On 06/24/2011 09:20 PM, Henry Baragar wrote:
Yes, I have a couple of workload distibution applications with highly
asymmertric workloads (see my message from from 2011-04-04 Re: LRU
broker queue in intuitive way on 3.0
because the desired granularity (single job) is so small
compared to all teh buffering involved, this kind of thing is, i think,
best handled by REQ/REP initiated from the worker.
i have handled similiar things before in this way, and even had
the central job dispenser solve for optimal job
Hi Henry,
OK. AFAIU what would solve your problem is to strictly limit the number
of messages in flight to a particular peer, ie. honour HWM exactly
without having excess messages in TCP buffers, right?
Right.
This have been in a pipeline for a long time but there were no
Interesting idea... hooking up a client/server model backwards implements
the pull architecure I need for my work distribution applications. It means
that the ventilator has to do its own queuing, but I think I can handle that.
I am going to give it a try.
Thanks!
Henry
On June 24, 2011
On June 24, 2011 03:59:41 PM Martin Sustrik wrote:
Hi Henry,
OK. AFAIU what would solve your problem is to strictly limit the
number
of messages in flight to a particular peer, ie. honour HWM exactly
without having excess messages in TCP buffers, right?
Right.
henry,
i'm not sure i followed all your examples, but i detected a general
meme which others
have complained of in the past. i admit to being sensitised to this issue
because i have worked
in this area for over a decade now.
the meme is that of conflating message routing with
26 matches
Mail list logo