The review of the papers of actuality Popular dance / weekly J2EE offered by Xebia.

Actuality editors / SSII

• Yahoo! distribute Hadoop
• JBOSS and eXo merges their solutions of Portlets
• Overhaul of JBOSS MESSAGING

RIA

• AjaxSwing 2.4

The corner of technology

• SoapUI 3.0 gets ready
• Comparison Gigaspace Coherence
• The dynamic conjuration for Popular dance becomes a reality

Ev nements of our community in France and abroad

• Return on CloudCamp

Actuality editors / SSII

Yahoo! distribute Hadoop

Yahoo! the availability of their own distribution of plan Hadoop announced. Answering a request of the community which wanted to have access to modifications performed by Yahoo! inside, this distribution is based on the last version of plan available Hadoop to foundation Apache.

With 13 committers on 22 that counts in total Hadoop, Yahoo! this plan hits widely by which popularity grows fast. In a context where Google, his direct rival, asserts his presence strongly in the field of distributed calculation and on the very coveting although appearing market of Cloud Computing, Hadoop plan constitutes a nice technological window for Yahoo!.

Wanting to cut short in the unavoidable accusations of try of fork or takeover of plan which did not hesitate to arrive, Yahoo! definite on the homepage of its distribution that all the patchs brought by this last were already contributed in Apache but did not appear in a stable version yet.

JBOSS and eXo merges their solutions of Portlets

JBOSS the constitution of a partnership with eXo which is going to give rise to fusion of to them solutions of respective Portlets: JBOSS PORTAL and eXo PORTAL. It is rather logically the solution of JBOSS that will survive this fusion.

In practice, the resultant portal will offer following specificity to the users:

• The expertise of JBOSS on its engine of Portlet which was more successful than solution of eXo in term of performance and of modularit .
• Interface user of eXo PORTAL as well as the comfortable possibilities of administration which he brought.
• The support of Portlets 2.0 such as defined by JSR-286. Both portals have already supported this finalised specification one year ago.
• Both firms will a priori be in even to offer the support on JBOSS PORTAL

This information was relieved by Touilleur, and Janus Boye offers an analysis of the impacts of this fusion for the current clients of one of both existent solutions.

Overhaul of JBOSS MESSAGING

JBOSS MESSAGING is an important brick of the deck undertaken by JBOSS. Having undertaken an ambitious overhaul of his server of application, the editor iterates operation with his middleware of messages. JBOSS MESSAGING 2.0 has indeed just been available in beta version. Tim Fox, project lead of Messaging, adds that this new version reuses his precursor only 5 % codes of. Among novelties, they will note especially:

• Embarquable in any type of application and integrable easily in JBOSS MICROCONTAINER or in other framework of injection of dependency (Spring?)
• Competitive stocking of information of journalisation
• Stocking JDBC leaning on Hibernate
• Support of the messages of big size
• Engine NIO supporting TCP, SSL, HTTP and APR

It is interesting to note that JBOSS MESSAGING 2.0 will lean on an engine NIO based on . This choice is unexpected, in measures him where JBOSS publishes , a rival framework of Mina created one year ago by Trustin Lee, then committer on Mina plan. The framework NIO of Apache benefiting from a bigger maturity and from a support APR will have been probably right of corporatism.

Moreover, the support of AMQP, STOMP and REST should arrive from the finished version 2.0 according to Tim Fox. These technologies of transport constitute important strategical attraction: AMQP has a strong potential as standard protocol of transport of messages M2M, STOMP assures compatibility with other middlewares of messages, and REST allows to be in a line with ActiveMQ and to give a Web access Service to MOM not without reminding of Amazon Simple Tail Service.

RIA

AjaxSwing 2.4

A plan which makes a lot of noise at present, with between others a presentation / d mo in Popular dance One, it is . To sum up, a plan of type RAP will be very similar to a plan RCP, to a big difference that the target will not be the desktop but the browser. It means that they continue developing in SWT LIKE (RWT) with all competences acquired for the development of a heavy client but it means that a light client will be got in the final.

But what is he for Swing? It is on the side of to TSS where they learn the release 2.4 of a plan making the bridge between Swing and HTML: AjaxSwing.

The tool will convert your Swing application into application HTML / Javascript in the runtime (contrary to GWT who makes it in compilation). One and particular knowledge asks by AjaxSwing for no specific modification: code is in pure Popular dance / Swing. That of happiness Rest supporting to see on a plan of wingspan if everything is indeed made without blow of particular screwdriver

For novelties of this 2.4, they note:

• Improvement memory (emptying of the buffer between requests)
• Better support of mouse listeners
• Improvement of components (FileOpenDialog, JTable cells )
• Better performances (regexp, communication server )
• Several corrections of bugs

Several d mos is with applications in Popular dance Starting Web and the same in Ajax. And of course a link towards the inevitable pages: licences and downloading.

The corner of technology

SoapUI 3.0 gets ready

Eviware relieves via The ServerSide, the announcement of the exit in version 3.0beta of SoapUI, the graphic tool to manipulate Web services.
After the addition of service management REST, this new opus goes on on the same way to make their manipulations easier still:

• From different requests and answers, the tool generates a diagramme XML of validation / WADL of fluxes and the same if exchanged contents are not XML (for example JSON or HTML)
• Addition of the support of operations OPTIONS and TRACE
• Rise of version of specification WADL.
• As for the import-export ratio of test of the operations of WSDL, it is possible now to have the same calculation on resources, methods and parametres REST (accessible only for version pro).

Besides, this new version includes following functionality between others:

• Support of Javascript: henceforth, besides groovy, him possible to use of Javascript as language of script
• The tool WS-I used to calculate its level of compliance WS-I BASIC Profiles is now packag with SoapUI
• The navigator of component is improved to allow the vision of answer HTML and REST in complex contents (ex: FLASH, SVG, Pictures )

To this is added the correction of bugs (to see the realease notes complete). SoapUI positions even more as the inevitable tool to manipulate Web services and REST services.
Soon on our blog a series of article introducing the main functionality of the tool.

Comparison Gigaspace Coherence

Gojko Adzic (via TheServerSide) offers on the blog an interesting comparison between two of the key products of DataGrid, Oracle Coherence and GIGASPACE XAP (damage to have left of quoted IBM EXTREM SCALE).
This comparison concerns five points:

• Mask of data: draw, the functionality of both products is quasi identical.
• Processing: GIGASPACE has a model of Processing orientated events and gives one good integration of Spring Transaction. Coherence is more limited.
• Deployment: GIGASPACE positions in return for of servers of applications, and has therefore comparable mechanisms of deployment, what Coherence does not offer. Of other one quoted, oracle's product gives a mechanism of s rialisation who authorises several version of the same class to be displayed on a grid.
• Scaling: Coherence gives a dynamic Scaling (in increase as in fall) where GIGASPACE works only with a number of knots data.
• Persistence: there still, given functionality is very comparable

And as usual in this type of comparison, no victor, no loser, only two products which answer different needs:

• GIGASPACE XAP, more capable to replace servers of applications and to accomplish distributed calculation. The model of Processing is more flexible and richer than that of Coherence
• Coherence, very competitive in the passive grids (reading only), because he SCALE dynamically, gives a flexible topology of data and has rich bookshops customers.

To sum up, this article offers a very good grid of assistant to decision.

The dynamic conjuration for Popular dance becomes a reality

R mi Forax, one of 4 committers of ASM, a framework of manipulation of bytecode widely used in the world Popular dance, has just announced that their last version supported education invokedynamic.

This education is defined by JSR-292 and aims at simplifying the conjuration of methods not known during compilation, what constitutes one of the main capacities of dynamic languages such as groovy. One waits therefore for these languages for a significant benefit of performance, since they will not have then any more to lean on an expensive mechanism of introspection or of dynamic generation of bytecode of conjuration.

Beyond dynamic languages which are the first addressees of this JSR, of many frameworks will be able also to use this possibility. Indeed, the manipulation of JavaBeans and of annotations, two major concepts in the world current JEE, lead to many conjurations of methods discovered in the runtime. In broad plans, it drives currently to the massive generation of bytecode to avoid very slow introspections, leading to a strong consumption of PermGen of JVMS.

The support of education invokedynamic by ASM, combined with the fact that the current draught of JSR-292 is already implemented and available in the preview of OpenJDK 7 since b59, allows to be optimistic as for the next revision of this JSR.

Ev nements of our community in France and abroad

Return on CloudCamp

CloudCamp was held in Paris on June 11th, uniting some actors of Cloud Computing. CloudCamps was initiated by Reuven Cohen and Dave Nielsen and aims knowledge and experiments on this appearing sector has share.

This CloudCamp turned out to be more commercial in the contents than JavaCamps: many partners so introduced their solutions during about one hour before real sessions (to the number of four, taking place in parallel) begin. Of each of them, they will keep:

• Architecture of Cloud: a solution of Cloud Computing must bring abstraction, elasticit and automation. If first two points are already covered by current solutions, the automation is not on the other hand another reality. This point will constitute an element determining for the adoption of Cloud Computing.
• Cloud and Security: the problems of security that they know in today IF new challenges reside with Cloud Computing, but are brewing in the case of clouds accessible publicly and divided across Internet without passing willy-nilly by devoted VPN.
• Cloud and Integration: the official report is dark. Plans thrown by the different editors and suppliers going surfing on the wave of Cloud Computing follow no standardization. To change supplier or solution on a plan constitutes currently a major modification.
• Aspects business of Cloud: the emergence of Cloud Computing drains the creation of new needs to be endowed. 3 levels of abstraction of Cloud Computing where are SaaS (Software package Ace A Service), Paas (Platform As A Service) and IaaS (Facilities Ace A Service) bring each the specificity in terms of mon tisation and from opportunities.