<<For most of the last 30 years, applications have dominated the
design center. We architected solutions around applications, hey we
even architected business processes around applications. Applications
have been the center of the software universe and everything has
revolved around them. However with the advent of firstly the Internet
and secondly the Web Service standards a whole new thinking is
starting to pervade. The radical thought is Copernican in nature.
Applications are in fact just planets in a solar system. The real
design center is the network and applications revolve around it.
This Copernican inversion is finally beginning to have profound
impacts all across the board. One of the main reasons Oracle is
interested in its Fusion middleware strategy, is that this middleware
underpins their whole next generation of applications. Salesforce.com
has given us a first glimpse of what a network oriented application
might look like, but many more examples are emerging now which will
move beyond what salesforce.com does today. The whole thrust of the
now famous memo from Microsoft's Ray Ozzie was on how this new network
centric computing model will change life for Microsoft. This
Copernican inversion is underway and its impact will be felt for the
decade to come.
So, what exactly are the major changes that are going to come about
and what will the landscape be like in 10 years time? I'm going to
attempt to answer that and make some calls about what will win out. In
doing that, I will also point out how Cape Clear is evolving its
product strategy to align with that 10 year vision. There are 5 major
themes which I see emerging
1. SOA -- The driver to simplicity and openness: There are a lot of
definitions about what is and isn't a SOA. I've written a lot about
this earlier and I don't intend to repeat it here. However, what
customers are expecting from SOA is a simpler, easier way to create
the software services that match their business needs. Core to this is
the support for the open industry Web Services Stack. This point has
now been accepted by all, and we are now seeing mature Web services
platforms like Cape Clear that enable simplified, declarative
programming models largely enabled by the underlying open standards.
Talking to some analysts recently, they remarked how every vendor is
now talking about how their SOA products make life simpler and easier
for the customer -- or they will do, when they eventually get around
to shipping them. I guess that imitation is indeed the sincerest form
of flattery, even if it only ships in the next release!
2. SOA meets Wiki, RSS and AJAX: There has be an awful lot of
nonsense spoken about the whole area of SOA Governance and SOA
repositories. There is a cluster of old time vendors who still seek to
cast all things SOA in the concrete technologies of the past. The
"classic" approach to the SOA repository has been to envisage it as an
old style relational database. This is just wrong and it misses one of
the key features of SOA.
SOA is on the border between the formal and informal worlds of
computing. On the implementation side (or the inward facing side) of a
SOA is the usual cluster of enterprise technologies: transactions,
security, reliability and so on. On the outward face are all the
informal, loosely-coupled aspects of a service. Its self describing
nature, its natural integration with Internet technologies, its high
level description of business services.
While governance affects the entire application life-cycle, the
critical issues relate to the challenges highly distributed
development imposes on human-to-human communication during
design-time. We believe that the right tools to design and manage all
the human aspects of SOA look more like a Wiki than a formal database.
Real people need to build and manage SOAs and a service should make as
much sense to a business user as it should to a developer. The Wiki
metaphor (and indeed its implementation) is the perfect vehicle for
sharing and managing SOA artifacts across an organisation. Then if
anyone makes a change to a service that you are interested in, RSS
should be able to inform you when that happens and you can make any
necessary changes.
3. The Inter-galatic App Server: This point could also have been
entitled "The Demise of J2EE". There are two reasons why we feel SOA
will cause the whole J2EE environment to recede as a power house.
First, the whole relevance of an execution environment for a
single programming language is questionable within the SOA world. As
the debate moves on to new center of gravity around services, then the
issue becomes one of how to assemble and manage those services -- not
how any given one of those services is coded. Secondly, whatever you
may feel about J2EE, most people now agree that the EJB model is just
way too complex. Projects like Spring point the future, such as it is,
for Java execution environments.
J2EE and its successors will be important in the same way that
CICS is, but they will not be center stage.
Instead we will see the emergence of a SOA Platform (or an
intergalactic app-server). This platform will be focused on solving
the problems of assembling different services together in order to
create a complete SOA application. This platform will not be focused
on a single programming language, rather it will cater for language
neutral services. This is a platform into which you can deploy and
manage an entire SOA application. This platform will give you full
control over all the policies associated with any given SOA
application. In short, it is a platform that covers the entire
life-cycle of a SOA application: From business requirements through to
governance of deployed services. This SOA platform is the grand
software prize of the next decade.
4. Subscription based everything: This is a short point as I can't
think of many new things to add to it. The movement to the new network
centric model of the software has made the subscription model
possible. People can now pay for services they use (over the network)
and pay for updates they receive from the network.
Subscription based pricing is inevitable as it reflects how
customers want to pay for software. Its a bit like digital music --
now that its caught on, it's unstoppable. Many large vendors are going
to push back on this one, believing that they have such dominant
positions with the customers that they can impose whatever pricing
model they like. All that will happen to them is that customers will
figure out how to do without them. Subscription pricing is one of two
key threats to large incumbent vendors: they are being boxed in by the
innovation of smaller vendors (and Open Source, see below), and they
are being hollowed out by subscription pricing.
5. The Open Source Paradox: The final major trend over the next
decade is Open Source. In order to explain the impact this trend will
have I need to rename it, as the term 'Open Source' is confusing. Many
people still confuse Open Source with "Free". What current Open Source
business models have proven is that access to the source code has no
impact on the commercial model of software. Most Open Source vendors
use a subscription based pricing model -- as it reflects the fact that
customers are very happy to pay for software as they use it. The
availability (or not) of the source code is a matter of indifference
to the value that most customers get from it.
So, the new name I propose for Open Source is "Participative
Software". The key thing to Participative Software is that it allows
individual groups of developers to participate openly in the creation
of software. The modular architecture required to allow for this type
of development inherently results in a more modular and extensible end
solution. As you look at research done on Participative Software, you
see that most of the projects do open copies of existing products -
app servers, office suites, operating systems, etc. - because this
helps to commoditize the key parts of the architecture with which
developers must integrate. However, these projects and products now
have compelling advantages: they are cheaper to buy and own than
proprietary alternatives, they are often more reliable and
trustworthy, and they are more flexible and extensible. The last point
will prove to be incredibly important within the heterogeneous world
of SOA. Of course, much of the software developed in this way can
still gain from additional integration testing, services, training,
support, and innovation around ease-of-use to address a broader
market. This is the added value you should expect from commercial
products based on Participative Software.
So over the next decade, most of the key elements in the
software ecosystem will be products based on Participative Software.
The greater community of open developers, combined with a compelling
subscription-based business model for adding the necessary value to
Participative Software will win out.>>
You can read these interesting thoughts at:
http://www.capeclear.com/annrai/archives/2006/03/leadership.html
Gervas
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/service-orientated-architecture/
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
