<?xml version="1.0" encoding="UTF-8" ?>
<rss version="2.0" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:wikidot="http://www.wikidot.com/rss-namespace">

	<channel>
		<title>Comments for page &quot;Location-independent Addressing with Messaging Middleware&quot;</title>
		<link>http://250bpm.com/forum/t-603811/location-independent-addressing-with-messaging-middleware</link>
		<description>Posts in the discussion thread &quot;Location-independent Addressing with Messaging Middleware&quot;</description>
				<copyright></copyright>
		<lastBuildDate>Sat, 01 Aug 2015 21:49:28 +0000</lastBuildDate>
		
					<item>
				<guid>http://250bpm.com/forum/t-603811#post-1931889</guid>
				<title>(no title)</title>
				<link>http://250bpm.com/forum/t-603811/location-independent-addressing-with-messaging-middleware#post-1931889</link>
				<description></description>
				<pubDate>Thu, 09 Jan 2014 23:41:24 +0000</pubDate>
				<wikidot:authorName>Holger</wikidot:authorName>								<content:encoded>
					<![CDATA[
						 <p>Hi,</p> <p>After working a bit more with message systems I would like to bring up the ROUTER functionality again. I totally agree still with DSN or SRV record solutions, but also like the idea of having a topology where I can send a message to a address which is not directly resolvable by DNS but routed by a given ROUTER which feels responsible. a pure DNS to IP mapping for for MyApplication name would be enough here? I would like to send a message to myApplication even if it not direct reachable by a TCP connection as it routed by a edge DEVICE, ROUTER etc.</p> 
				 	]]>
				</content:encoded>							</item>
					<item>
				<guid>http://250bpm.com/forum/t-603811#post-1679489</guid>
				<title>(no title)</title>
				<link>http://250bpm.com/forum/t-603811/location-independent-addressing-with-messaging-middleware#post-1679489</link>
				<description></description>
				<pubDate>Sun, 06 Jan 2013 18:48:17 +0000</pubDate>
				<wikidot:authorName>Martin Sustrik</wikidot:authorName>								<content:encoded>
					<![CDATA[
						 <p>The main point of the article is that the location-independent addressing is no business for messaging. It should be done on other layer(s) of the stack.</p> <p>As for the exact mechanism to use, I am in no way expert in the area. AFAIK there are mutliple attempts for solution, but none of them is widely accepted (to the point where you can just do connect (s, &quot;myApplication&quot;) and the stack will do the rest for you).</p> <p>Not being an expert means that I can only guess what are the reasons behind functionality so fundamental not being ubiquitous. Anyway, my guess would be that the problem is the inherent tension between requirements for LANs (dynamic discovery of services) and requirements for WANs (statically configured service names, such as DNS SRV records).</p> 
				 	]]>
				</content:encoded>							</item>
					<item>
				<guid>http://250bpm.com/forum/t-603811#post-1678540</guid>
				<title>(no title)</title>
				<link>http://250bpm.com/forum/t-603811/location-independent-addressing-with-messaging-middleware#post-1678540</link>
				<description></description>
				<pubDate>Sat, 05 Jan 2013 10:03:26 +0000</pubDate>
				<wikidot:authorName>Holger</wikidot:authorName>								<content:encoded>
					<![CDATA[
						 <p>I strongly agree with DNS name resolution here, but what you describe sounds so familiar with running DIAMTER over SCTP. Where SCTP gives you some form of message absctraction and DIAMTER Base Protocol can (it is not mandatory) resolve the endpoints by asking for SRV/NAPTR records described in RFC 3588 and updated recently here:<br /> RFC 6733 section 5.2 (sorry cant post a link here)</p> 
				 	]]>
				</content:encoded>							</item>
					<item>
				<guid>http://250bpm.com/forum/t-603811#post-1656294</guid>
				<title>(no title)</title>
				<link>http://250bpm.com/forum/t-603811/location-independent-addressing-with-messaging-middleware#post-1656294</link>
				<description></description>
				<pubDate>Mon, 17 Dec 2012 20:13:55 +0000</pubDate>
				<wikidot:authorName>martin_sustrik</wikidot:authorName>				<wikidot:authorUserId>939</wikidot:authorUserId>				<content:encoded>
					<![CDATA[
						 <p>SRV records are step in the right direction. However, there's much more to it. On LAN, for example, you want the services to be registered and discovered dynamically, rather than inserted by hand into DNS (think DNS-SD) Also, I guess, some kind of scoping is needed (local service with same name shadows the global service) Etc.</p> <p>Once the following works out of the box on LAN, we're pretty close:</p> <div class="code"> <pre> <code>s1 = socket (AF_INET, SOCK_STREAM, 0); bind (s1, &quot;myApplication1&quot;); ... s2 = socket (AF_INET, SOCK_STREAM, 0); connect (s2, &quot;myApplication1&quot;);</code> </pre></div> 
				 	]]>
				</content:encoded>							</item>
					<item>
				<guid>http://250bpm.com/forum/t-603811#post-1655936</guid>
				<title>(no title)</title>
				<link>http://250bpm.com/forum/t-603811/location-independent-addressing-with-messaging-middleware#post-1655936</link>
				<description></description>
				<pubDate>Mon, 17 Dec 2012 15:52:55 +0000</pubDate>
				<wikidot:authorName>Mordy Ovits</wikidot:authorName>								<content:encoded>
					<![CDATA[
						 <p><em>&quot;And once again, the solution is pretty straightforward. We need a directory service to map application names to IP addresses.&quot;</em></p> <p>Isn't what you're describing the same as DNS SRV records? It sure seems like it:<br /> hxxp:// en.wikipedia.org /wiki/ SRV_record</p> 
				 	]]>
				</content:encoded>							</item>
				</channel>
</rss>