It’s been a few years since JavaOne felt like a fun, vibrant conference. But this year, thanks to the release of the OpenJDK and Java SE 7, attendees had plenty to talk about and learn at the show.
But Java wasn’t the only focus of this double conference that shuts down large parts of San Francisco for the week. Oracle’s announcements were highly developer related as well. With new databases, new machines, and a reprieve for the death sentence that was expected for SPARC, Oracle OpenWorld and JavaOne were chock full of cool tools.
Oracle has long been the company to which all roads in an enterprise lead. With databases that underpin many of the applications inside Fortune 500 companies, information is just used to living inside Oracle systems. But what’s been happening over the last two years has made that paradigm troublesome for the folks on the cutting edge of analytics. Perhaps the biggest time sink in analytics today revolves around the ingress of information into Hadoop.
With terabytes or even petabytes of unstructured data out there to contend with, just moving the info in an Oracle database into a Hadoop cluster for analysis could take an entire day. What’s the solution? Why, selling an Oracle-branded Hadoop system. Oracle’s retinue of business intelligence analytics will run on these Exalytics machines as well. But the real excitement is that these systems are built for fast data transfers and instant query replies, two things that can cause a lot of pain for Hadoop users. If Oracle can continue to innovate around Exalytics, you can bet it’ll be a very compelling force in the Hadoop world.
The Oracle NoSQL Database
You may better know this particular database by its older names: Berkeley DB or Sleepycat. This XML database has always been, essentially, a key value store. A highly performant key value store.
Oracle’s been quietly selling this in-memory database for years, but only at its conference this week did it categorize the database as a NoSQL. And the company did this very begrudgingly. It would seem that the term “NoSQL” is seen as slightly negative for Oracle, as it has lost some business to the myriad competitors in this new market for highly scalable and fast database alternatives.
No matter whether Oracle likes the term or not, it’s now playing in that market. It tweaked Sleepycat to be able to play in this environment by expanding the scaling capabilities of the database, and by offering it as an advanced caching layer for its database applications.
Another technology that seemingly got lost in the Oracle/Sun shuffle, JavaFX is the long-awaited UI scripting layer for Java. It’s targeted as an RIA development platform and includes many features that are still only dreamed of in standard Java, such as built-in media streaming support and links into graphics accelerators.
But for users, JavaFX is probably the best thing to happen to Java since SWT. It’s a flashier, snazzier interface layer for Java applications, which have always lacked a bit of panache when compared to native applications. And for the zealots among you, JavaFX will soon be open-source technology as well when it’s pushed into the OpenJDK 8 specification. If all goes well, JavaFX 3.0 will be part of OpenJDK 8 and Java SE 8.
The expanding PaaS market
Everyone and their mother now offers a Java PaaS. From the “old man” of the platform-as-a-service world (Heroku), to highly enterprise-focused GigaSpaces, to CloudBees’ build-based approach, to Jelastic’s “any Java” strategy, there’s no lack of innovation in the growing PaaS market.
But these third-party companies aren’t the only ones supporting Java in their clouds. Oracle announced its own cloud offerings at the conference, and they look to address many of the issues that exist in other clouds. Oracle is hoping to have an open cloud, without vendor lock-in, which is almost the exact opposite of its database strategy. Additionally, Oracle claims any Oracle database can be moved into its cloud. If that proves to be true, this could become a very compelling cloud offering for enterprises struggling with scaling out their infrastructure.