Tuxedo and Application Grid
By Ruma Sanyal on Apr 13, 2009
Since the BEA acquisition there may have been questions regarding the future of Tuxedo at Oracle. I guess when a product has been around as long as Tuxedo has, this is probably reasonable. After all, Tuxedo – a middleware product – has been around lot longer than the term middleware itself! As the product marketing director for Tuxedo at Oracle, I want to put those questions to rest. Tuxedo is very much a strategic product within the Oracle Fusion Middleware portfolio. It’s thriving with a strong roadmap and with an extremely loyal customer base.
The more interesting question is, what is Tuxedo’s strategic fit within Oracle Fusion Middleware? Tuxedo is in many ways the original ‘application grid’ platform. (For more information on application grid, click on the link to watch this video http://blogs.oracle.com/applicationgrid/2009/04/video_whiteboard_on_applicatio.html.) As the requirements of an application grid architecture demand, Tuxedo provides a runtime infrastructure with mainframe-class quality of service. Tuxedo provides rock-solid reliability with its inherently distributed nature and transactional integrity with its to phase commit support. It provides fives nines availability and linear scalability to extreme levels. In terms of performance, currently we are working with a customer that has shown 667K tps in a benchmark study with Tuxedo! Pretty impressive.
Think of Tuxedo as the application server for applications developed in languages other than Java. Applications developed in C/C++ that may be running today on a homegrown app server are well-disposed to benefit from a robust offering like Tuxedo. According to a report by Evans Data Corp even in 2007, 31% of programmers identified themselves as primarily C/C++ programmers. Not counting the ratio of C/C++ applications that are already running in various data centers, this means that going forward, quite a few applications will be developed on C/C++, and the ratio of C/C++ applications will remain significant.
As far as applications running on mainframes, the Tuxedo value proposition is simple, but extremely compelling: Customers are spending between $3M-$5M for 1000 MIPS for their mainframe applications. Many of these applications written in COBOL can be rehosted to run on Tuxedo. This will result in tremendous cost savings, reduce dependency on legacy infrastructure, and reduce dependency on a graying workforce. Read the whitepaper on mainframe rehosting with Oracle Tuxedo here http://www.oracle.com/webapps/dialogue/dlgpage.jsp?p_ext=Y&p_dlg_id=6739015&src=6642146&Act=7."
The more recent Tuxedo releases have focused heavily on SOA. One of the prime intents here is to expose the IT asset rehosted on Tuxedo to become a first-class SOA participant. More on this in my next blog. Meanwhile, if you want to know further details, please listen to the webinar – Lower total cost of ownership for mission-critical SOA here http://www.oracle.com/webapps/dialogue/dlgpage.jsp?p_ext=Y&p_dlg_id=6739015&src=6642146&Act=7."