MOM and SOA on Linux? 48
dogmeant asks: "I have yet to see MOM (message oriented middleware) offerings on Linux much less open source projects that address the same. And while we're at it, what about tools that address SOA/BPM (Service oriented architecture/business process modeling) type architectures on Linux? Is the Open Source community ready to take on challenges like this?" If anyone else out there has this particular itch, maybe this will be another niche in which Open Source software can catch some Enterprise mindshare.
Buzzword bingo! (Score:3, Funny)
Ah, middleware (Score:4, Insightful)
It works well for companies because:
1) They can afford to put time and investment into 'the latest fad', whereas in the OS community the time is better spent elsewhere. (The point isn't phrased very well sorry - read on)
2) It is good for companies to produce middleware because they can offer a high level of intergration between their products - hence locking you into their whole range. E.g. Lotus domino - news server,web server, database all in one. In the OS community you would chose the best database applicable, the best web server applicable, and quickly write some scripts to communicate between them.
3) Middleware solves problems that people don't know they have - so don't notice that itch to solve them. However in business, companies produce such software, then tell other companies they have the problem, and solve it. (I'm not very good at getting point across today - sorry.) For example: IBM do a message passing program that basically consolidates errors between servers. Say a hub breaks down, and suddenly a load of machines report they can't access the machines behind the hub. What you want is for the computers to talk to a central node, and the central node to reduce all the errors down to a "x,y,z machines can't be reached, they are all behind router r, hence I suspect r is down." Then when the machine come back, delete the errors.
Re:Ah, middleware (Score:3, Insightful)
Very true. Some are solving the problem in very complicated way, like Staffware, we used it in writing workflow application for JSP. It's supposed to make life easier, but turn out the total effort to collaborate and manintain this damn Staffware server is multiple times more than it takes to write a workflow mechanism on our own from the ground up.
What's worse is that most middleware introduce a vulnerable layer in a supposingly robust system. The above-mentioned system has a AIX+Oracle backend and it's supposed to be very reliable, but most of the time the system cease to function when that Windows server which host Staffware goes down. I swear to God I do not make this up.
When they sell you middleware and say "slip this in and you can save the money doing this layer yourself!". They lied, they just want your cash.
May be that's the major reason why opensource community seldom target on middleware, as most of them are useless.
Re:Ah, middleware (Score:2)
Oracle's 'reference platform's right now are Sun and/or Linux. If you must stay with oracle, you might want to switch. If you must stay with AIX, you might want to go db2 (though the first will be far easier... oracle export -> oracle import and you're done)
Re:Ah, middleware (Score:3, Interesting)
I must admit the choice is based on the business needs, not on the technical side, but technically I don't think DB2 shine on AIX. DB2 is great on MVS, I was a DB2/MVS developer.
Regardlessly, this combo seldom has problem. So far the AIX down once, and Oracle never. That Staffware on Windows 2K, on the other hand, down twice per week, at least.
It goes deeper than that. (Score:5, Insightful)
On the other hand, it seems that most open source developers are motivated to scratch their own itch. They are drawn to apps that are cool and apps that meet their needs. This is perfectly understandable but, it does result in a derth of open source business apps that are either mundane or just plain boaring. For this reason, open source offers an emormous assortment of "toy" apps. There are countless cd burners/rippers, MP3/Ogg players/encoders, Linux utilities and so forth. There are also a ridiculous amount of frameworks and libraries, usually all repeating the same theme. If I see another "yet another framework" project I'm going to be sick.
There are only a few business type open source projects and even fewer ones that could be considered mature enough and good enough to run your business on them. If a business wants a particular app they must build it themselves. This of course costs money and in some cases it costs a huge amount of money. And there in lies the problem. Most businesses that have spent large amounts of money on development are not interested or willing to give it away by making their project open source. For these two reasons there are very few business type open source apps like middleware and it will be a long time before there are, if ever.
Re:Ah, middleware (Score:2, Interesting)
Building and especially maintaining infrastructure software is way to expensive for a single company to take on. This is where Open Source comes in. Albeit, a consortium of some sort would have to be put together.
Re:Ah, middleware (Score:1)
Re:Ah, middleware (Score:2)
Imagine you had two servers (er lets say mysql and apache). They have a very loose coupling between them. They can talk to each other (you can serve web pages from a database) but you wouldn't really say that they were joined together. With most implementations you could probably easily rip out mysql and replace it with postgres or something with fairly minimal fuss.
Now a company, if they owned both, would put far more resources in coupling them - creating a strong coupling. Apache would have very specific mysql code, and the config files would have mysql settings in it, and it would bring up mysql if it wasn't running, and there would be a single tool to configure both - sort of like Lotus Domino style.
It requires a lot more resources to maintain a tight coupling then it does a loose coupling. The apache people would have to re-consider every so often the pro's and con's of being coupled with mysql instead of say postgres, and change a lot of stuff if they decide to switch, forcing everyone else to switch along with them.
Yet again I'm not really making myself clear - I'm sorry. One last really small example. Scripting. KDE apps don't have a place to put in scripts - like in MS word with it's VB - because they don't need. They just export a load of functions through dcop, and you can write the scripts in your favourite editor, with your favourite language (if it has the bindings).
Re:Ah, middleware (Score:1)
Kid: "No thanks..."
Open Source middleware already exists (Score:2)
Middleware exists because people who "quickly write some scripts to communicate between" are basically amateurs who create systems and applications, which look very much like a pile of knotted wool. Getting useful business information out of such systems is like attempting to unknot said pile of wool. Design and architect are not words which are part of their vocabulary.
Middleware removes the complexity, it turns X*Y levels of complexity into X+Y. If you want a quick idea of how middleware helps, this bloke's done a web page which gives an idea:
http://www.archeus.plus.com/colin/middlewa
Choose your poison...
NNTP, IRC, SMTP if you want to roll your own out of components not specifically designed for the purpose.
Bond, XMLBlaster, Nirvana, Jabber and the rest, for open source systems designed for the job.
And the commercial boys all support Linux now, so if you want some accountability, features and support, MQseries, webmethods etc are also available.
I suspect the reason you're having problems getting your point across because you're wrong in almost every respect.
Hope this Helps.
isectd (Score:4, Informative)
IBM has middleware on Linux (Score:4, Informative)
Vinci [ibm.com] is a SOA for Linux. The site describes it as "a local area service-oriented architecture designed for rapid development and management of robust web applications"
I have attended presentations on MQ series, and found it interesting, but never had any opportunity to use it. I found the Vinci paper by googling.
Re:IBM has middleware on Linux (Score:2)
Backend CIS system on a mainframe, and a java based linux webserver need to talk? No problem.
IIS/MSSQL box need to pass data back to DB2 running on an AS/400? No problem.
Sun/Oracle box talking to a custom homegrown app? No problem.
It's a unified API, and a unified 'paradigm' of communication. It can talk TCP, SNA, and every other network type out there. It runs on every os/hardware platform you're likely to see in production. And it offer's a promise of message delivery. It won't just try and give up. It guarantees (sp) that the message will get to where it's destined to go.
The downside is that it's kinda heavy weight. It's also a little network chatty, and it's also rather expensive. However it does make it easier for your developers. If they all know how to get from their application to a Message Queue, then all of your developers can get data back and forth between ANY of your applications.
All of that said, there is some talk in the jabber community about trying to move into the 'middleware' space. That'd be a really interesting thing to watch. Doing that guaranteed delivery isn't something to be taken lightly, and if they could crack the portability problem (how many of us have as/400's to play with?) it'd be pretty cool.
MQ (Score:2, Informative)
It's neither open source of free of charge, but certainly not expensive.
Re:MQ (Score:1, Offtopic)
Java... (Score:1, Informative)
The first that springs to mind is XMLBlaster. Google is your friend.
Re:Maybe I've been living in the... (Score:1)
Unless of course, you think Object Oriented programming is bullshit, and you code in FORTRAN.
Jabber, XMLBlaster and friends (Score:5, Informative)
Jabber [jabber.org] is normally thought of as yet another IM system, but "Jabber is an open XML protocol for the real-time exchange of messages and presence between any two points on the Internet" (from the Jabber site). Its first application has been IM, but it is by no means limited to IM. Jabber is a protocol specification, and there are several open and closed source implementations of clients and servers.
xmlBlaster [xmlblaster.org] is a more traditional MOM offering, under LGPL. It supports numerous protocols and bindings into a number of languages.
As usual, Google offers a lot of advice on the topic of 'mom middleware "open source"', including a list of MOM implementations [middleware.org] which tells us that JORAM [objectweb.org] is also open source, and an article entitle Open Source in Middleware [ebizq.net].
Re:Jabber, XMLBlaster and friends (Score:3, Interesting)
We're using Jabber on Linux to support http://cougaar.org/ administration - to stop/start/configure nodes, check on status, etc, etc. It's pretty sweet, especially when used via Rich Kilmer's Jabber4R Ruby Jabber client. Check out http://www.infoether.com/ruby/jabber4r/ for more info...
Tom
Find unused Java code with http://pmd.sf.net/
AQ (Score:3, Informative)
If you have Oracle on Linux you can use AQ, and others have mentioned IBM's MQSeries. Also there is a pure Java one called SonicMQ but I've never used it.
Re:AQ (Score:1)
Doesn't sonicMQ use a database though?
I ran into this when of my customers set up an Oracle database to use sonicMQ and I was wondering why they didn't just use Oracle AQ instead.
MOM is SOL on Linux? (Score:4, Funny)
The joram project (Score:4, Informative)
The project is spear-headed by INRIA [inrialpes.fr] (a French research institute). The whole system is open-source and they are doing quite advanced stuff (including group communications).
JBOSSMQ (Score:2, Informative)
No, no, no (Score:2)
Mom is SOL on linux.
At least mine is. Your mom's an astronaut.
SOAP frameworks (Score:1)
The biggest advantage that Cordys gave us, is that it has several "application connectors" that make it a breeze to access data in old propietary ERP systems like Baan and SAP.
I hope they're not paying you for research.. (Score:2)
http://www.google.com/search?q=mqseries+linux
http://www.google.com/searchq=message+oriented+
Honestly
Re:I hope they're not paying you for research.. (Score:1)
Haven't seen spread mentioned (Score:2, Interesting)
MOM/SOA/ESB on Linux (Score:3, Interesting)
While I'm a big fan of open-sourced solutions, there are plenty of commercial MOM offerings out there that run on Linux. A really solid one that I've worked with is SonicMQ. It runs on Linux and many other platforms [sonicsoftware.com]. Being built 100% in Java makes that possible.
As for SOAs, there's another offering from Sonic called SonicXQ that offers a standards-based SOA that includes support for web services, content-based routing, transformation, and itinerary-based process flow, all done using proven standards. They recently added a suite of XML tools from the acquisition of eXcelon that gives them sophisticated XML storage and handling as well as stateful conversational BPM.
The industry has labeled this new form of SOA that combines MOM, Web services, cbr and transformation as an Enterprise Service Bus (ESB). Lots of talk about the ESB being a more flexible cost-effective integration strategy than the traditional integration brokers approach.
And this too runs on Linux.
You might want to do a little research... (Score:4, Informative)
jboss (Score:1)
Plenty of MOM's to choose from (Score:1)
Re:Plenty of MOM's to choose from (Score:1)
Re:Weren't these called TPM's before (Score:2)