Inquiry: redesign of init property lookup

December 22nd, 2011

We are asking for community input on the following: over time, our initialization mechanism (property file, system properties, properties supplied programmatically) has been polluted a bit and as a result the lookups are not always intuitive. We would like to improve this and are working on a new design proposed here. Let us know what you think!

Proposed new property lookup procedure when Atomikos starts:

1. default properties are looked up via a property file in the Atomikos jars (should always be found)
2. override with specific custom properties looked up in the file in the classpath (like now)
3. if no found: NO logging to System.err any more (so no need to disable that with system props either)
4. override with specific properties that are set programmatically on the UserTransactionService
5. resolve any placeholder properties (i.e., ant-style expansion of references to other props)
6. this yields the final properties to use for the init procedure

What do you think? Did we miss something important?


Why BRIC countries love Atomikos

September 15th, 2011

BRIC countries have the tremendous opportunity to start out with no legacy, no old-style infrastructure like application servers, no message brokers, no enterprise service bus (ESB) or other nonsense like that - the typical stuff big enterprise vendors will try to trick you into (if they do: just ignore them - you will be better off).

Instead, BRIC countries have the luxury of being able to pick a light-weight model such as JEE without application server to achieve unlimited scalability at virtually no cost, using commodity hardware and software.

With the recent economic crises in both the US and Europe, it is the BRIC countries that are helping out (instead of the other way around). The future will involve a lot more BRIC. And with that, a lot more Atomikos.

Western companies that want to be as agile, productive and competitive as the BRIC countries: what are you waiting for? Call us to get an introduction and migration path for your legacy.

Coming soon: Replacing Tuxedo with Atomikos

September 15th, 2011

If you have been purchasing the ‘Enterprise Architecture’ portfolio from the big O then guess what: you’ve paid way too much for something you don’t need, and which will overload your (way too heavy) hardware even more.

We’ve been claiming this for years so it is about time to show it: how to replace Tuxedo applications with open source and free software. Coming soon to this site!

Stay with us…

PS Tuxedo is a registered trademark owned by Oracle

Announcing PostgreSQL support via 2ndQuadrant

June 30th, 2011

In a recent post we’ve announced our end-to-end support program. We are glad to be able to add PostgreSQL to the list of supported JDBC vendors - thanks to and via our PostgreSQL support partner 2ndQuadrant.

Another success story: avoiding the complexity of the appserver

June 27th, 2011

This blog post describes yet again how you can use best-of-breed components to do the job.

In particular, it describes integration of Atomikos as well as all of the other cool libraries that together outperform the application server in simplicity, cost and performance… Good job!

Atomikos survey results

June 13th, 2011

I would like to dedicate this post to all of our users who have been kind enough to give us their feedback via our online survey (you know, the one you receive by email once you register for download).

First of all, thank you very much for your feedback - this type of information is invaluable for any business, especially a start-up like us. Being agile as we are, any inconveniences found can be quickly corrected in our products and our business model.

So what are the results so far? Well, I won’t keep you in suspense any longer so let me just outline the summary:

  • About 90% of the respondents would feel bad about missing our products in future projects
  • About 60% have recommended our products to other people

Up to you to decide if this is good or bad - but we certainly are pleased with that result! Once again, a big thank you to our community!

Cloud: the dawn of the SOA era

June 8th, 2011

The recent announcement of Apple’s cloud illustrates yet again that a new era is about to start: that of the cloud - and with that, the massive adoption of service oriented architectures (SOA).

This will have an impact so huge that it is hard to underestimate. Some of the obvious changes that will happen are most likely the following:

  • Infrastructure (operations) will become agile - at the expense of IT jobs in operations. Applications will be deployed with unprecedented ease and speed. This is because the cloud infrastructure will transparently allocate the resources that now are allocated by human administrators.
  • SOA will become mainstream. This is because even the most common infrastructure services will be virtualized, running in the same or a different VM from the application, and scaling up and down as needed.

The interesting part - at least to us - is that transactions will become more distributed than ever. We are ready for this change - are you?

Announcing end-to-end support

June 6th, 2011

We are going through exciting times here at Atomikos! Today, we are pleased to announce that we will be offering single-point-of-contact, end-to-end support across different back-end systems for the following database vendors:

  • Oracle
  • DB2
  • SQLServer
  • Sybase
  • Informix
  • PostgreSQL via 2ndQuadrant (added on June 30 2011)

…and the following queuing vendors:

  • MQSeries
  • ActiveMQ
  • SonicMQ

This is an initial list that will probably grow over time (stay in touch to find out).

What does this mean? Well, it means that Atomikos can offer support from the queue to the database and back again without you having to worry about dealing with different support channels.

This support is optional and not included in the standard support subscription. Contact us to find out more…

How to leverage Atomikos in the cloud

June 4th, 2011

Data consistency in the cloud is a totally different ball game - as perfectly illustrated by this newsletter issue from ZapThink. Essentially, you have to cope with some form of partition tolerance and stale data values in caches. This is usually summarized as the CAP issue (for more details, see the ZapThink post).

Luckily, the solution to the problem can be as simple as outlined in our solution to the CAP problem.

Life is good :-)

Reflections on the cloud

June 1st, 2011

The cloud is an interesting phenomenon, to say the least. Just last week, I witnessed a demo by Adrian Colyer at the What’s Next (Paris) event that left me very impressed because I could easily understand the problems that it solves:

  • Slow infrastructure setup is replaced by instant availability of resources (databases, hardware)
  • Scalability is trivially easy: on command is all it takes
  • Services are promoted as the means of reuse - effectively enabling SOA strategies

I know from experience that these are all pain points in today’s IT infrastructure for just about every big organization. So for mainstream operations, I think the cloud is going to be a killer.

While this is all fine, I still wonder how this will work out for one of our core markets: finance. Sure enough, initial offerings are showing up, but so far nothing really interesting for our retail banking audience. Add to this the fact that most banks are typically very conservative (they have to be) and the cloud for the banking world seems a bit more of a future situation than for the rest of the industry.

But certainly that time will come too, and for those days Atomikos has the exact kind of technology for cloud-enabling transaction guarantees that banks depend on so much. That is because we offer the third generation of TP monitoring technology: light-weight, embeddable, perfectly scalable and SOA capable too.

The future is bright;-)