Talk:Enterprise application integration
|This is the talk page for discussing improvements to the Enterprise application integration article.|
|WikiProject Computing||(Rated C-class, Low-importance)|
Merging and stuff
User:Michael Hardy put a redirect from 'Enterprise Application Integration' (capitalised) to this page, after he spotted the duplication. However he didn't do any merging, so I've been rescuing content from this other page, and trying to merge it into this article. As the same time I tried to make sure the whole article flows and makes sense. So it's looking good now, but that last sentence mentioning 'businesss process' is out on its own a bit. Need to expand upon this aspect of EAI. -- Nojer2 13:44, 13 Dec 2004 (UTC)
I would like some comments in regards to what is missing in Enterprise_application_integration software that is generally addressed by Enterprise_service_bus software, in companies that do not have J2EE setup. 126.96.36.199 00:05, 3 October 2005 (UTC)
Very nice! I'm new to the Wiki world, but I was glad to see an article like this out there when I looked. I do EAI for a living, and have a few thoughts on this topic. I would like to do a bit of expansion on the article. Is this allowable?
188.8.131.52 12:23, 11 March 2006 (UTC)
- Certainly, that's even very welcome, so be bold. Maybe the Wikipedia:Manual of Style is of help in doing so. Welcome! --S.K. 11:26, 13 March 2006 (UTC)
That's excellent, and thanks for the style reference to look at! I'll see what I can do with it. The article needs a few illustrive diagrams, and some drill down into the topic. Let's see what I can do with it :) -- 184.108.40.206 12:14, 16 March 2006 (UTC)
Looks like I did an okay job of cleaning it up. Anyone else care to take a stab, or is that good enough? --Disavian 03:22, 25 April 2006 (UTC)
I slightly de-emphasized the ESB POV. Added some more generalized notions of EAI architecture. Despite the current ESB-craze there are lots of other solutions out there- the SIF EAI standard being a notable example. Alex Jackl 00:50, 16 October 2006 (UTC)
There's a bit of an inconsistency with the ESB article that ought to be clarified. Namely, this EAI article says EAI can have a hub or a bus architecture (with "bus" linking to the ESB article). But the ESB article says, "Contrary to the more classical enterprise application integration (link to EAI article) approach of a monolithic stack in a hub and spoke architecture, ..." implying that EAI does not / cannot use an ESB. Maybe the inconsistency is due to the above change; maybe it's due to lack of consensus on what EAI and ESB are, in which case the problem can't yet be solved. Huttarl 7 Nov. 2006
- That's not what the sentence is supposed to imply. It could certainly be worded better. EAI is an umbrella term, encompasing both the bus and hub approach (and any other approaches you can think of) to application integration. The sentence is saying the "more classical" approach to EAI is be hub approach. -- Nojer2 10:10, 20 December 2006 (UTC)
I'd like to add a company under the Commercial Solutions section but it would be an external site link. The other companies listed there all have wikipedia pages so they only link internally. Does a company need a wikipedia page to be listed?
The 70% failure statistic is not well supported by the cited article. The article says that EAI projects suffer from a roughly 70% failure rate, but cites no studies/evidence to support this. Without stronger evidence, shouldn't this entry use less definite wording? Maybe "it has been suggested that..."? —Preceding unsigned comment added by 220.127.116.11 (talk) 20:09, 10 September 2009 (UTC)
I propose to remove the lists of Commerical and Open Source solutions. They add nothing to the article content, are mostly out of date, and if required could very well form their own list article. Thoughts? --HighKing (talk) 17:47, 27 April 2010 (UTC)
- I also propose removing the products under the Future section. The MS product is not an EAI product and is in such marginal use in its field (BPM) that listing it first is actively misleading. I support creating a list in a separate article that can be discussed and maintained properly.
Under section "Improving connectivity", the paragraph referring to "fully meshed point-to-point connections" is completely absurd. This statement (along with the formula) is implying that every application used by an enterprise is connected to every other application in the enterprise. This is scenario is unrealistically pessimistic, and I challenge anyone to prove me otherwise by citing a real-life instance of this problem. From my industry experience, I have found that the number of connections is approximately equal to the number of points. Please provide a citation to these facts or remove the paragraph. Thank you. — Preceding unsigned comment added by Jonez176 (talk • contribs) 15:30, 5 March 2011 (UTC)
- I don't see why the paragraph is "absurd". Sure, it's pretty clearly a worst-case scenario, but it's been qualified by limiting itself to "fully meshed point-to-point connections". This scenario is not unusual when considering subsets of the full suite of applications present in an enterprise.
- However, there are numerous citations that could be given in support, and I happen to agree that one or more citations is necessary to supports the stated claims in the remainder of the section.
- With respect to your comment, do you have anything to support your claim that the number of connections is approximately equal to the number of points? I'd expect something like the Watts and Strogatz model to be applicable, which offers some support to your comment. Redazzo (talk) 11:37, 30 July 2012 (UTC)