Every few years, a new computing language comes along to save the world. Not long ago the messiah was Java, a language that has grabbed some important niches but has not noticeably eroded Microsoft’s hegemony in the network industry. Utopian notions of Java as the “new network platform” now seem hopelessly old-fashioned.
All eyes are now on XML, a fresh canvas onto which we’re mapping our fondest hopes for cross-platform application interoperability. XML stands a better chance than Java of loosening Microsoft’s proprietary sway over most tiers of distributed networking. The ongoing evolution of XML standards is beyond the control of any one platform vendor, a claim that proponents of Sun’s Java cannot make.
XML and related specifications, such as XML Namespaces and XML Schema, are well on the road to widespread enterprise deployment in the next three to five years. Adoption is always the sincerest form of flattery, and vendors are implementing or have implemented XML in their core product architectures.
However, no one seriously regards XML as the basis for a new operating environment, in the traditional sense of that term. Rather, XML is just a very adaptable language for defining higher-level markup “vocabularies” for various application domains. Communicating applications need not share a common development model, net protocol, operating system, database or programming language as long as they can exchange and interpret documents, messages and other objects formatted in XML. Senders and recipients of XML-formatted objects are free to process the objects as they wish, without being tightly bound to one another.
XML renders traditional platforms irrelevant. It provides a versatile language for decoupling distributed applications from their operating environments. Vendor-dominated operating environments will take a subordinate role to cross-platform services, such as directories, Web publishing and electronic commerce. And these cross-platform services will increasingly implement XML down to their cores.
Most fundamentally, XML is helping shift industry momentum away from tightly coupled computing models and toward messaging-oriented middleware (MOM). XML’s emergence is contributing to a broad decline in platforms’ reliance on remote procedure calls and other inter-application communication schemes, such as Distributed Component Object Model (DCOM) and Common Object Request Broker Architecture (CORBA), that bind distributed objects tightly to one another.
DCOM and CORBA will still have a role in this new order, primarily in supporting distributed applications that demand tightly coupled components in relatively homogeneous environments, such as intranets. Even here, XML will be pivotal, with software vendors increasingly relying on XML “wrapper” technology to integrate otherwise incompatible object technologies across multivendor, multiplatform environments.
There is a clear trend toward loosely coupled, platform-independent distributed applications that rely on XML metadata. We can already see the forerunners of this trend in extranet application products under development at companies such as Hewlett-Packard and Silknet. These vendors are using net directories to store and manage the complete XML-encoded workflow business logic of distributed applications, using XML-based metadata to keep tabs on database fields, security profiles and other entities. One might refer to such an application framework as an XML-based service bus. This approach received a boost from Yahoo’s recent announcement that it will use HP’s e-speak technology as the basis for future e-commerce services.
Integration of XML with Lightweight Directory Access Protocol Version 3.0 directory infrastructures, under the Directory Services Markup Language initiative, should only accelerate the trend toward using XML syntax to define directory-resident metadata pertaining to workflows, services and applications.
This all just goes to show that XML will be like oxygen in the networked environment of the 21st century: vital, pervasive and taken for granted.