﻿<?xml version="1.0" encoding="utf-8"?><rss version="2.0"><channel><title>Ayende @ Rahien</title><link>http://ayende.com</link><description>Ayende @ Rahien</description><copyright>Copyright (C) Ayende Rahien  2004 - 2021 (c) 2026</copyright><ttl>60</ttl><item><title>Corey commented on Silverlight Queues: Design</title><description>Spending some time reading up on WebSockets I think that could give some interesting options over PollingDuplex. 
</description><link>http://ayende.com/4568/silverlight-queues-design#comment14</link><guid>http://ayende.com/4568/silverlight-queues-design#comment14</guid><pubDate>Wed, 04 Aug 2010 14:19:31 GMT</pubDate></item><item><title>Chris Chilvers commented on Silverlight Queues: Design</title><description>I think this solves more than just Silverlight, such as MSMQ isn't that great if the client's end point keeps changing, such as if you have a laptop out in the field connected using 3G. The client won't have a constant connection (perfect reason for using queues), but also won't have a consistent IP or DNS name.
  
  
I'd name the server side per-client holding area mailboxes, since they work similar to say POP3 email (or using Comet, IMAP with push).
  
  
The main points I saw is the mailboxes need to be cheap, as you'll probably need to create a large number of them as you'll need one per client, which might translate to one per session. From what I've seen a queue per client for MSMQ wouldn't be pretty, not sure about Rhino.Queues or ActiveMQ. RabbitMQ seems to eat memory when it has a large number of queues from what I've read.
  
  
I was thinking something as simple as serialize each message to the file system server side, perfect if the message order doesn't matter as you can use the message's guid as the file name. If the serialized format matches the wire format then it's ready to go, which can also work well with systems like AtomPub.
</description><link>http://ayende.com/4568/silverlight-queues-design#comment13</link><guid>http://ayende.com/4568/silverlight-queues-design#comment13</guid><pubDate>Tue, 03 Aug 2010 18:01:34 GMT</pubDate></item><item><title>hrvoje commented on Silverlight Queues: Design</title><description>I would love to see some framework/toolkit for that!!
  
  
Now i'm thinking on using rhino bus or some other bus framework, and polling duplex services for Silverlight. Then, on silverlight side, another messaging pub/sub system (mvvmlight) for internal SL needs. Unified system would make this much easier. 
</description><link>http://ayende.com/4568/silverlight-queues-design#comment12</link><guid>http://ayende.com/4568/silverlight-queues-design#comment12</guid><pubDate>Mon, 02 Aug 2010 15:29:16 GMT</pubDate></item><item><title>Tom Allard commented on Silverlight Queues: Design</title><description>We are currently developing a Silverlight app that communicates with a webserver that is using rhino servicebus to orchestrate messaging with different services. Currently we are using httppollingduplexbinding to forward incoming messages server side over wcf to caliburn.reactive where messages are pushed to the relevant components client side. 
  
  
What we think would be nice is simply to retain the rsb api client side. Unfortunately we have had no time to investigate in how to do that ourselves. Offline availability is not a requirement for us, so we don't really have a need for having persistent messages client side.
</description><link>http://ayende.com/4568/silverlight-queues-design#comment11</link><guid>http://ayende.com/4568/silverlight-queues-design#comment11</guid><pubDate>Mon, 02 Aug 2010 09:30:48 GMT</pubDate></item><item><title>Ayende Rahien commented on Silverlight Queues: Design</title><description>[www.gliffy.com/gliffy/#templateId=blank&amp;signup=1](http://www.gliffy.com/gliffy/#templateId=blank&amp;signup=1)</description><link>http://ayende.com/4568/silverlight-queues-design#comment10</link><guid>http://ayende.com/4568/silverlight-queues-design#comment10</guid><pubDate>Mon, 02 Aug 2010 08:59:17 GMT</pubDate></item><item><title>Robert Vukovic commented on Silverlight Queues: Design</title><description>Off-topic
  
  
in what program do you make those nice diagrams and where do you find such nice pictures of server, computer, firewall ... ?
</description><link>http://ayende.com/4568/silverlight-queues-design#comment9</link><guid>http://ayende.com/4568/silverlight-queues-design#comment9</guid><pubDate>Mon, 02 Aug 2010 08:57:31 GMT</pubDate></item><item><title>Corey commented on Silverlight Queues: Design</title><description>I agree polling duplex is the right way to go here for silverlight. I do think there's  a need. It would be very nice for those trying to achieve a CQRS solution in Silverlight. Could also be something that sets RSB apart from other solutions if there is an easy bridge between the two.
</description><link>http://ayende.com/4568/silverlight-queues-design#comment8</link><guid>http://ayende.com/4568/silverlight-queues-design#comment8</guid><pubDate>Sun, 01 Aug 2010 20:20:17 GMT</pubDate></item><item><title>Ayende Rahien commented on Silverlight Queues: Design</title><description>Alexandria uses Rhino Queues, which can't be hosted in IIS, but you can use MSMQ (or any other queuing tech) as well
</description><link>http://ayende.com/4568/silverlight-queues-design#comment7</link><guid>http://ayende.com/4568/silverlight-queues-design#comment7</guid><pubDate>Sat, 31 Jul 2010 14:44:30 GMT</pubDate></item><item><title>Tuna Toksoz commented on Silverlight Queues: Design</title><description>Silverlight Polling Duplex might make things a bit easier.
</description><link>http://ayende.com/4568/silverlight-queues-design#comment6</link><guid>http://ayende.com/4568/silverlight-queues-design#comment6</guid><pubDate>Sat, 31 Jul 2010 14:41:31 GMT</pubDate></item><item><title>Frank Quednau commented on Silverlight Queues: Design</title><description>Just looked through Alexandria. Interesting! I am just wondering - what protocol does rhino.bus use? can it be hosted in IIS much like a WCF service?
  
  
Just questioning myself whether message request/response + batching could be put on top/into WCF...
  
  
Websync looks quite elegant, as opposed to ye old' polling. The other day, a colleague of mine explored a Comet approach based on this article: 
[www.codeproject.com/KB/aspnet/CometGrid.aspx](http://www.codeproject.com/KB/aspnet/CometGrid.aspx)  
  
First things first ;) Do we have an open implementation of Comet for the .NET stack?
</description><link>http://ayende.com/4568/silverlight-queues-design#comment5</link><guid>http://ayende.com/4568/silverlight-queues-design#comment5</guid><pubDate>Sat, 31 Jul 2010 14:37:14 GMT</pubDate></item><item><title>Ayende Rahien commented on Silverlight Queues: Design</title><description>Frans,
  
The problem that this solves is the same problem outlined here:
  
[msdn.microsoft.com/en-us/magazine/ff796225.aspx](http://msdn.microsoft.com/en-us/magazine/ff796225.aspx)  
  
The need exists, witness only by the fact that people asked me about it.
  
</description><link>http://ayende.com/4568/silverlight-queues-design#comment4</link><guid>http://ayende.com/4568/silverlight-queues-design#comment4</guid><pubDate>Sat, 31 Jul 2010 09:31:16 GMT</pubDate></item><item><title>Ayende Rahien commented on Silverlight Queues: Design</title><description>Howard,
  
Probably, that is highly dependant on the requirements that you have for ttl for messages, I think.
</description><link>http://ayende.com/4568/silverlight-queues-design#comment3</link><guid>http://ayende.com/4568/silverlight-queues-design#comment3</guid><pubDate>Sat, 31 Jul 2010 09:30:15 GMT</pubDate></item><item><title>Frans Bouma commented on Silverlight Queues: Design</title><description>"How much interest is there in Queuing system in Silverlight?"
  
  
Isn't the question: "what problem would this solve?", and the answers to that should be analyzed and if the best answer involves queueing, then one should look at a queueing system but IMHO it now looks like you're looking for a problem which fits a solution which was written without knowing the problem.
</description><link>http://ayende.com/4568/silverlight-queues-design#comment2</link><guid>http://ayende.com/4568/silverlight-queues-design#comment2</guid><pubDate>Sat, 31 Jul 2010 09:26:47 GMT</pubDate></item><item><title>Howard van Rooijen commented on Silverlight Queues: Design</title><description>Could WebSockets offer an alternative mechanism for broadcast / polling?
</description><link>http://ayende.com/4568/silverlight-queues-design#comment1</link><guid>http://ayende.com/4568/silverlight-queues-design#comment1</guid><pubDate>Sat, 31 Jul 2010 09:19:28 GMT</pubDate></item></channel></rss>