|
  Exclusive Interview with
Eric Newcomer
CTO,
IONA Technologies .
Read the 2nd interview [July 2005] for more on ESB and the new
Celtix
IONA Technologies [www.iona.com],
with revenues of $68M in 2004, was a pioneer in standards-based
software integration with its Orbix family of products.
Founded in 1991, on the campus of Trinity College, Dublin
(Ireland), IONA, now with over 350 employees worldwide, prides
itself as a leader in enterprise IT integration – spending the last
decade tackling some of the world’s toughest enterprise IT
integration projects, most notably in the financial and
telecommunication sectors. IONA sets out to drive down the cost and
complexity of enterprise IT through a synergistic blend of superior
technical expertise and incisive software offerings. Thus,
according to IONA, 70% of the financial services organizations in
the Fortune Global 100 rely on IONA's software, as do 80% of the
world’s largest telecommunications companies. With credentials such
as these IONA is a unique position to play a pivotal role in shaping
the future of SOA-based solution.
IONA’s Enterprise Service Bus (ESB) product family, which
runs natively on all major integration platforms of import including
IBM mainframes, is known as Artix. IONA’s ESB offering, easily
extensible thanks to a standards-based architecture that encourages
‘plug-ins’, is meant to adroitly address the manifold challenges of
complex, mission-critical IT integration projects. With Artix 3.0,
the latest version of the company’s extensible ESB, IONA intends to
dominate the enterprise SOA arena – particularly when it comes to
customers with IBM mainframe applications.
Eric Newcomer, the CTO of IONA, is here interviewed by
Anura Guruge, the editor at
large for IT In-Depth to find out how and why Artix differs from
other offerings on the market – and to get some insights as to how
IONA feels XML Web services-based IT integration vis-à-vis what is
possible with CORBA and Orbix.
ERIC NEWCOMER
In his role as
Chief Technology Officer at IONA, Mr. Newcomer is responsible for
directing and communicating IONA's technology roadmap, as well as
IONA's product strategy as it relates to standards adoption,
architecture, and design.
Mr. Newcomer
joined IONA in November 1999, after nearly 16 years at
Digital/Compaq, where he held a variety of technical and management
roles. He joined IONA as the company's transaction processing
architect, and also served as IONA's Vice President of Engineering,
Web Services Integration Products.
Mr. Newcomer
leads IONA's participation in all standardization activities, and
has been involved in Web services standardization activities from
the beginning. He was a founding member of the XML Protocols Working
Group at W3C, which produced SOAP 1.2. He is also a former editor of
the Web Services Architecture specification at the W3C and is IONA's
primary representative to OASIS and WS-I. Mr. Newcomer is co-author
and editor of the recently published Web Services Composite
Application Framework (WS-CAF) set of specifications and co-chair of
the WS-CAF technical committee at OASIS.
A frequent
speaker at industry and company events and contributor to popular
journals and Web sites, Mr. Newcomer is the author of the
best-selling Understanding Web Services (published in May
2002 by Addison-Wesley), and co-author with Phil Bernstein of
Principles of Transaction Processing (published in January 1997
by Morgan Kaufman). His new book, Understanding SOA with Web
Services, written with Greg Lomow, was published by Addison
Wesley in December 2004.
Q1: Eric,
with Artix, the company’s extensible ESB, IONA indeed appears to
have a very competent and promising host integration solution, with
uncompromised support for XML Web services as well as CORBA. Can
you please tell us what you consider to be the true value
proposition of Artix, especially in terms of other competitive
mainframe-based offerings?
A1:
It is true that Artix offers a compelling host integration solution.
It’s also important to remember that because of the product’s broad
platform support and its ability to extend the enterprise qualities
of service inherent in mission-critical systems into new
applications, Artix is ideally suited to address the most complex
integration challenges facing the enterprise, independent of IT
environment. But since your readership is most interested in what’s
going on in the mainframe space, I’ll try to focus my comments and
talk about the Artix value proposition in that context.
Clearly, the idea of host integration is not new, and companies have
been trying a variety of different methods to make host-based
applications more flexible and accessible to the enterprise. Screen
scraping and other similar approaches to mainframe integration
allowed companies to use host systems in new ways while preserving
performance, but these approaches often proved inflexible and
brittle. The advent of application servers and message queuing
middleware improved flexibility, but added new complexity and often
suffered from performance problems.
With Artix for z/OS and OS/390, IONA provides a means for companies
that have made significant investments in mainframe technology to
truly service enable the applications running in this environment.
IONA offers a solution that can run both on, or off the host and
expose mainframe transactions as standard Web services. By creating
intelligent service endpoints that can support a range of
transaction types, including those exposed directly via IMS and CICS,
(using COMMAREA for example) and CICS through 3270/BMS, Artix
addresses both the performance and flexibility shortcomings
associated with other integration approaches. Further, the plug-in
architecture unique to Artix supports changes to transports or
security models without requiring changes to the application, thus
extending its usefulness to the enterprise, and doing so in an
efficient and cost-effective manner.
Q2:
Eric,
I get a feeling that much of the inherent power of Artix (especially
Artix for z/OS) lies in your highly optimized ‘adapter’ technology
for extracting CICS and IMS transactions and data. Is that really
the case and is this a key feature that allows you to be so
confident about enterprise-class performance and scalability?’
A2:
What you refer to as ‘adapters’ we tend to refer to as ‘plug-ins.’
And you are right to say that it is the Artix plug-ins, and the
architecture that supports these plug-ins, that allow us to speak
with such confidence about enterprise-class performance and
scalability of service-enabled mainframe systems. I’ve already
talked about how the plug-ins support IMS and CICS transactions, but
that’s only a part of why we believe this approach is the most
advantageous..
When many mainframe shops hear the words ‘Web services’ used in
conjunction with integration, they’re skeptical. In their minds,
this equates directly to SOAP over HTTP, which due to issues of
performance and scalability is often simply not an option. Unlike
those limited Web services products, Artix natively supports
multiple message formats and transports. This multi-format,
multi-transport support is a direct result of Artix’s plug-in
architecture. What may be ever more interesting, is that in the
situation where a mainframe shop may have proprietary or home grown
transports or message formats to support, Artix plug-ins can be
easily created to meet those requirements.
This ability of Artix to take advantage of the transports and
message formats already in use in the enterprise, and to extend the
enterprise qualities of service associated with an organization’s
mainframe applications is why we are so confident that Artix
delivers the enterprise-class performance and scalability required
by our customers and prospects.
Q3: So
Eric, what about security and what hooks do you provide with
mainframe-based security offerings such as RACF and ACF2 for access
control and authentication?
A3:
Security is absolutely something that falls within the enterprise
qualities of service that we take very seriously. Out of the box,
Artix allows the end user to securely integrate SSL and Kerberos
with RACF, ACF2 and Top Secret. This seamless integration of
security models ensures that new Web services-based applications can
be accessed easily and securely without requiring reengineering of
the original host-based application.
In fact, we’ve seen this in action at a customer site where
call center representatives are able to sign-in on a .NET client and
have that authorization carry across to the host based applications
that deliver information back to the .NET client. Other security
technologies can easily be integrated as well by developing and
configuring new plug-ins.
Q4:
Eric, if we are talking just about CICS, and a customer wants to
expose some CICS functionality as XML Web services why should they
opt for Artix rather than just using IBM’s new Web services
capability with the z/OS
CICS Transaction Server V3.1?
A4:
We’d say that the option you describe does not offer the flexibility
that enterprises are looking for from initiatives to service-enable
the mainframe. It all goes back to the idea of extensibility.
Artix supports not only the usual transports like HTTP and MQ
Series, but also IIOP, and because of its plug-in architecture, can
be made to easily support other transports that may be found in the
enterprise. This means that by using Artix to expose CICS
transactions, you’re opening up the options for which non-host
applications can consume those services, without degradation of
performance or qualities of service. In other words, it’s not about
exposing just CICS functionality, but it’s really about what other
applications and systems can use that CICS functionality once it’s
exposed.
Q5: Eric,
does IONA, in the context of Artix, plan to support some form of
grid computing to further enhance the value-proposition of Artix?
A5:
While grid computing is an interesting subject, we plan to maintain
our focus on providing general solutions for the most challenging
and complex integration problems facing today’s enterprises.
Q6: Eric,
as my last question, given IONA’s unique experience with CORBA, how
would you compare and contrast the pros and cons of XML Web
services-based IT integration with what is possible with CORBA?
A6: The
goal of both approaches to integration, CORBA and Web services, are
very similar. CORBA’s popularity was, in many respects, a result of
CORBA’s ability to uncouple the interface from the underlying
logic. In a way it’s somewhat ironic to hear people talking about
Service-oriented Architecture as something ‘new’ when IONA can point
to successful SOA deployments in our customer base, SOAs built more
than a decade ago and built on CORBA. It’s also interesting to note
that we continue to see the adoption of CORBA in new markets and
geographies. CORBA is not going away anytime soon.
But what we are seeing in the industry now is a desire to
broaden the success of the CORBA approach using new technologies,
such as those based on Web services. Much like CORBA, Web services
provide formal, well-defined interfaces for integration, independent
of language, operating system or messaging middleware already
existing in the enterprise. Because Web services are based on XML,
however, they offer additional benefits of simplicity, broad
adoption, and data independence.
As we’ve addressed in earlier questions, Artix is proof
positive that in a mainframe environment, Web services can be used
to describe interfaces between clients and servers that ultimately
avoid the common brittleness associated with screen scraping or
wrapper approaches.
Thank you very much Eric.
This will definitely help
folks appreciate what IONA and Artix can do for them when it comes
to SOA-based IT integration.
Read
the 2nd interview [July 2005] for more on ESB and the new Celtix
|
