The cloud by itself is scalable, says Sven Hammar, CEO for Apica. But applications can run very differently on different cloud vendors’ platforms. And some of the same issues that tracked application hosting providers can surface in certain cloud settings. Take for example the ‘noisy neighbor.’  Cloud is meant to handle spiky phenomena such as Super Bowl marketing campaigns. But if your apps are co-located near that suddenly surging campaign, it can be a problem for your app’s performance.

”Sometimes you have a noisy neighbor. Adequate online performance monitoring can detect if you are located next to, for example, a busy campaign site,” said Hammar.  The questions that need to be answered are not unfamiliar to veterans. How many concurrent users can I handle? Where are my bottlenecks? What happens when you update the cache?

SpringSource’s Rod Johnson, whose Spring Framework rose to prominence as a kinder and gentler way to do Java ventured into this battle earlier this year when he admitted to OSGi’s complexity. As SpringSource’s OSGi dm Server is one of the poster children for OSGi success to this point, Johnson found he had to do some clarifying. The server is now part of the Eclipse Foundation portfolio. Here, per TheServerSide.com is Rod Johnson’s take on OSGi:

(a) OSGi is a great solution for complex applications with stringent modularity requirements;

(b) typical business applications (from which we make the bulk of our revenue) don’t have such requirements;

(c) our efforts to reduce the complexity of writing server-side OSGi applications were promising, but the road to simplification was longer and less certain than we’d hoped. Thus continuing down that road at the Eclipse Foundation, in partnership with other companies and individuals, was a natural move.

I found Bennioff’s assertions about the future of enterprise computing both entertaining and inspiring, but a little bit vague on the details. He was absolutely right about a lot of what he was telling the enterprise marketers to do. We need to be more open, more democratic, and better connected. I was really impressed by his vision for Toyota Friend, a private social network for owners of Toyota automobiles that lets the cars and their owners interact on a somewhat personal level. (Bennioff’s answer to IBM’s Volt?)

But how are those goals going to be realized? After the keynote, I got some answers from Salesforce VP Peter Coffee, head of platform research. ”Everyone knows that when the CEO asks what’s going on with the cloud, you better have an answer,” he told me, “Now, the same is going for social networking.” He cited marketing examples like Pepsi Refresh and Toyota Friend as potential differentiators that make social networking an imperative.

From the technical side, Coffee filled me in on three important points about the new Chatter system that’s enabling networks like Toyota Friend. First, he said that Chatter is not just a social networking app that can be deployed behind the firewall. According to Coffee, Chatter adds social networking functionality like feeds and profiles to existing enterprise applications.

Second, Coffee assured me that “Chatter is not a walled garden. Anything with an API can participate in Chatter interactions.” That means if your server monitoring software (for example) has an output API, it can potentially warn you proactively about common situations before they turn into big problems.

Thirdly, Coffee explained the new messaging model that Chatter is based on. According to Coffee, Chatter breaks down the barriers that currently exist in most organizations between documents and messages. Chatter has a content library with versioning management that employees can link to in an update similar to the way that they would on Facebook. But instead of going out to the entire organization, the update goes out to others who are following the project/customer/document/tags/whatever that are associated with that update. It’s supposed to make creating interoffice communications a similar experience to posting on Facebook or Twitter, which people already like to do.

Coffee claims that people demonstrably prefer this type of communication, when it’s properly implemented. He warns against allowing a social network to become just another silo and posits that one of the important keys for successful service integration is enabling abstract entities like accounts and documents to start conversation threads, using complex event processing in the background.

In February, the push on Wall Street was for systems that turbo-charge aggregation and pricing of financial transactions, order routing, algorithmic trading and market data management. These areas had looked like losing bets in late 2008, when banks and investment houses merged under tremendous financial pressures. But apparently the sleek technology represented by Complex Event Processing (CEP) helped more than a few of these firms recalibrate risk and begin to climb up the financial slope again.

In May we found enterprise architecture (EA) frameworks like The Open Group Architecture Framework (TOGAF) and the Zachman Framework offer powerful reference models through which enterprises can build out infrastructures. And yet Zachman’s dreams of a perfect world where enterprise architects all follow a common framework may be no more than a distant hope. Changing business needs, licensing costs, skill availability, M&A, corporate politics and many other factors are all aligned against it.

Randy Carey, a former director of information strategies, developed his own personal philosophy on EAI – buy as much pre-made as possible and then perform tailoring where needed. Likening the process to buying a suit, he says most people can’t afford and don’t need custom-made formal attire but instead can simply have a mass produced product tailored to fit. “In other words you need to figure out where on the cost-curve you want to be,” he said in July.

Newton’s Law seems at times to play out in the IT shop. Every action in software development does seem to create an equal reaction. These thoughts arose in August as we spoke with Shridar Mittal, CEO, ITKO. Mittal told us that, yes, SOA actually is getting adopted out there, that as a result applications are now very integrated and that the work on these apps is more often distributed across multiple teams.

When it comes to issues of cost-benefit analysis, many organizations turn to Gartner for help. The Gartner Magic Quadrant rates technology vendors on completeness of vision and ability to execute, and is much watched in software circles. It uses quadrants to graphically represent leaders in a specific niche and time frame. In November, we examined Gartner’s Magic Quadrants for Application Infrastructure — one for SOA-Style Application Projects and one for Systematic Application Integration Projects.

In December, we took another look at Enterprise Application Integration (EAI) There was a time when EAI was the antithesis of SOA, but EAI is making a bit of a comeback within the SOA firmament. The fact is that it never really went away. The funny thing is that SOA has made a lot of progress in generally integrating all kinds of diverse systems, most notably mainframes. But, in some instances, Web services wrapping of mainframes may be reaching an illogical conclusion. There remain cases where EAI is the more economical alternative.

With 2010 behind us, we’re all looking forward to 2011. No matter what this new year brings, rest assured that we here at SOA Talk will have something to say about it.

He sees a dichotomy in this field..

Fascinating to me are the different assumptions people make in the different worlds between the REST/HTTP “scale out” designs and the mainframe-derived “scale up” designs. This is likely to remain an area of continual evolution.

Oddly, or ironically, SearchSOA.com just ran a piece rolling up some various transaction monitoring events. Read ”End-to-end testing for SOA and enterprise transactions, or Show me the transactions.”

There are a lot of questions yet to be resolved in this ongoing story – What will happen to Java? Where for art thou, MySQL? Will Oracle become a hardware company? Whither Netbeans? And so on.

Eric Newcomer, in his role of OSGi champion, has an additional question: How will this merger affect OSGi? We have featured OSGi more than a bit on both this blog and on the SearchSOA.com site, and therefore are glad to share some of Eric’s thoughts on OSGi in the light of the proposed mega merger here. He suggests he may clear up some murkiness on the Sun side that previously seemed to be out and about.

He writes:

“Oracle has been a strong supporter of OSGi. They are well represented on the enterprise expert group that I co-chair, and one concrete action I expect is that they will move to eliminate Sun’s formerly schizophrenic attitude toward OSGi for Java modularity.

Sun had supported OSGi through its adoption in GlassFish, but had also competed with it in Project Jigsaw and to some extend in the JSR 294 work as well.  Among other things this created uncertainty over the role of OSGi in Java’s future, especially for enterprise applications.  Given Oracle’s strong participation in the OSGi enterprise effort to date I expect this acquisition will have a positive impact on the OSGi enterprise initiative.”

The OSGi Specification defines a dynamic module metadata system for Java and a service-oriented programming model with which the modules interact. The specification defines a registry for service lookup, and a collection of built-in services for common functions such as security, lifecycle management, and logging. The OSGi framework has been adopted by the Eclipse Foundation and by every major Java vendor as a platform on which to build and ship middleware products and open source projects, including application servers, enterprise service buses, and IDEs.

As the core platform has become widely adopted in products and open source projects, the OSGi Alliance began to receive requirements for more explicit support of enterprise applications. The OSGi Specification began its life as JSR 8, back in 1999, intended for use in home automation gateways. Since that time OSGi technology has achieved some level of adoption in various embedded applications for the automotive, mobile telephone, and home entertainment. By September, 2006, the OSGi Alliance had received sufficient indications of interest in an enterprise edition to hold a workshop to explore the possibility of chartering an enterprise expert group (EEG).

Since its first meeting in January, 2007, the EEG has spent the past two years creating detailed requirements and designs intended to better support enterprise Java applications. The work will result in a major update to the specification in mid-2009 (two prerelease drafts have been published) that extends core framework services and adapts existing enterprise Java technologies to the OSGi framework to meet enterprise application use cases. The major features include a mapping of the Spring Framework component model called the Blueprint Service, a mapping of existing distributed computing protocols to the OSGi service model, and mapping key parts of Java EE such as Web apps, JDBC, JPA, JMX, JTA, JNDI, and JAAS.

The industry has already embraced the benefits of OSGi-enabled modularity. Next is to improve its support for enterprise Java applications by adapting technologies already used in those applications. This goal is to help OSGi developers more easily create enterprise applications in a standard way.

Eric Newcomer is a distributed computing specialist and independent consultant. Newcomer is a chair of the OSGi Alliance Enterprise Expert Group and former CTO of IONA Technologies. He writes a blog on OSGi matters.

Moreover, he notes, the WS02 approach has merit because “the registry/repository has not been left as an afterthought.” Hailstone says most of the problems he sees with implementation of SOA these days seem to revolve around governance failure, rather than technology failure.

He continued:

Potentially, OSGI should make it easier for customers to mix ‘n match components from different vendors within their SOA environment. Whether there will actually be a demand for this is not yet obvious, but it provides a good safety net in that, should a competitive product have (for example) a more appropriate Event Processing engine it would not drive a large hole through the technology selection strategy. It should also make for a more dynamic ecosystem of specialist technology providers, reducing the amount of custom integration needed to make, for example, a policy enforcement engine integrate into a different platform.

Sanjiva Weerawarana said his company’s approach to architecture is focused on interoperability.

Said Weerawarana, WS02 co-founder and CEO: “One of the key tenets that we are focusing on is interoperability, not portability If the situation arises where we have to invent a custom API, okay. But we will do all that work in Apache. That provides protection. People can innovate without being constrained.”