Why 'wireless middleware' will be one of the most important
infrastructure investments you will make over the next three
years
There can be no e-business without mobility
The need to
mobilise key enterprise information assets - extending the
accessibility of information and applications to mobile workers -
has never been stronger. Two main trends are behind this:
- every company of significant size now relies on IT systems and
digital information to conduct its business effectively, both
internally and in partnership with others
- many company workers are no longer desk-bound, but increasingly
work from multiple locations, or are truly mobile.
These two trends create a major tension, because the trend towards
e-business has a centralising effect on information; and most
e-business systems are designed to be accessed from desktop PCs.
However, e-business also relies for its effectiveness on the
timeliness and quality of the information that is input. Where does
the input come from? It can come from interactions with customers
and prospects, from people who fix assets 'in the field', from
distributors and delivery companies, and so on.
In other words, the quality and timeliness of e-business input
comes from ill-defined and mobile locations outside the walls of
the enterprise. E-business is nothing without mobility of
information and applications, yet enterprise IT suppliers and many
consumers of these systems treat it as an expensive
exception.
Middleware: the glue for the wireless enterprise
Wireless
middleware provides the
 | | "Wireless application and
information services will be implemented from existing assets,
rather than via the creation of new ones" | | | | | |
| | Source: Ovum | | |
|
|
runtime infrastructure and development services necessary to
deliver digital content, software applications, and e-services to
users of mobile computing devices.
Why do you need wireless middleware? The answer is simple.
Today there is considerable diversity in devices themselves, the
network services they attach to, and the wireless data protocols
they use to send and receive information. In ten years there may be
one global public network standard, and one wireless data protocol.
| | "The most capable devices and
networks are the ones that are the least widely adopted" | | | | | |
|  | Source: Ovum | | |
|
|
However at the very least, there will continue to be many different
types of device - not because of technology constraints, but
because different people playing different roles in different
situations will demand form factors that suit them.
Reducing costs and complexity
Attempting to deliver a
data service to a significant population of workers, by directly
addressing each type of device and network that might be used, is
therefore very difficult and likely to be prohibitively expensive.
Wireless middleware hides the complexity of delivering services to
different mobile computing devices by translating behind the scenes
between network protocols, application protocols, user interface
standards, security implementations, and so on. Wireless middleware
allows you to deliver sophisticated data services to mobile devices
cost-effectively.
Applications, applications, applications
In practice,
it is highly unlikely that the applications and information that
you will want to deliver to mobile workers will be completely new.
In most cases, wireless application and information services will
be implemented from existing assets, rather than through the
creation of new ones.
This fact has an important impact on wireless middleware
technology. Most companies will be extending, not creating,
applications and information, so the value of wireless middleware
products is principally in dealing with the complexities associated
with the variety of devices and networks that wireless services
must address. Unlike other types of middleware, such as application
servers and integration technologies, the value is not in dealing
with the complexities associated with hosting and managing large
and complex software applications.
Catering for different environments
All software
applications are automated ways of creating, sharing, analysing and
accessing digital information - but they do these things to
different extents. Applications, and the services which deliver
them, therefore vary in terms of the kind of environments that
users need in order to be able to find, understand, create and
manipulate information; and in the nature of the information
itself.
In the context of analysing the requirements that applications
place on the environments in which they run, there are two
extremes: communication-centric applications and process-centric
applications.
Communication-centric and process-centric
applications
Communication-centric applications include
applications such as voice communication, text messaging, unified
messaging, e-mail and collaborative 'groupware' applications.
The value of a communication-centric application is largely
dependent on those people able to connect to the application and
share information. If there are enough of the right type of people
'connected', the application becomes valuable. Conversely, the
sophistication of the user environment that hosts a
communication-centric application is of secondary interest. Change
in the implementation of a communication-centric application is
largely driven by the network environment. This might arise because
the application needs to reach another group of users using a
different kind of network or device, for example.
Process-centric applications include customer information and sales
applications; field service applications; and office productivity
applications.
The value of a process-centric application is dependent on the
quality of the environment through which information can be
visualised, created and manipulated. In contrast, the value of a
network that connects the device to a central information store may
be incidental to the overall value of the application. What is
more, process-centric applications are often characterised by the
fact that they need to provide users with functionality at all
times - even when there is no network connection. In contrast to
communication-centric applications, change in the implementation of
a process-centric application is largely driven by internal company
pressures - such as changes to business processes - rather than by
external pressures.
Different applications create different kinds of
complexity
The most capable devices and networks are the
ones that are the least widely adopted. The most widely-used
devices and networks will be far less capable. This much is basic
economics.
In addition, there is currently a great deal of heterogeneity in
today's mobile devices and networks, and this is not going to go
away.
These two facts together lead us to three uncomfortable
truths:
- if you want to deliver an application or service to the largest
possible audience, you will have to deal with large amounts of
heterogeneity in devices, networks and protocols
- if you want to deliver these applications you will also have to
settle for delivering lowest-common-denominator functionality,
because most devices and networks only provide limited processing
and transfer capacity
- if you want to deliver an application or service that is as
sophisticated as possible, you will have to strongly limit the
variety of devices and networks over which you deliver your
offering (and therefore the population of users).
Consequently, process-centric applications - which implement
specialised processing of information in order to create a rich or
highly structured user environment - cannot in general be delivered
cost-effectively to large populations of users. Conversely,
communications-centric applications, which aim to create
connections between members of an audience, cannot assume that
users will have access to sophisticated application delivery
environments
Different types of applications require different services from
wireless middleware technology, in order to deliver to end users
the kind of experiences they want. Despite vendors' claims to the
contrary, the suitability of a given wireless middleware solution
is not universal; it is dictated by the type of information or
functionality you want to deliver to users. Applications are not
just content.
At a high level, then, there are two sets of functions that are
important in wireless middleware platforms:
one to deal with the 'complexity in the network', which arises from
the need to
- deliver applications and services over multiple devices and
networks
- another to deal with 'complexity in the information', which
arises from the need to deliver usage environments that allow users
to interact with and manipulate information in sophisticated
ways.
Components of wireless middleware
By dealing with
complexity, and smoothing out the differences between multiple
machines, systems, networks and protocols, middleware products ease
the job of the software developers who have to extend existing
information and application assets out to mobile workers and
consumers over cellular networks.
Indeed, the true value of any middleware solution should be judged
by the degree to which it makes the job of software developers
easier. Wireless middleware products are no exception.
There are seven main areas of wireless middleware
functionality, all of which hide certain complexities from the
developers who are responsible for delivering wireless applications
and services:
Wireless middleware is important, but the market is
immature
You need to consider the role that wireless
middleware can play in helping you implement your e-business
strategy - but your ability to procure the best-fit solution relies
on you understanding the overall problem space, and convincing
suppliers to show you how they fit into it.
The wireless middleware market in 2001 is immature, but growing
fast. Despite the overall industry conditions, new players (both
start-ups and established) are entering this market all the time,
and there seems to be enough room for any player with an ability to
differentiate itself. Unfortunately, few of these companies are
making it clear how what they do helps address real requirements.
The market is set to change over the next year. By mid-2002, the
enterprise middleware players such as IBM, Oracle, BEA, Microsoft
and Sun/iPlanet will make serious and concerted moves into this
market. Until now, they have not made serious investments - despite
some of their marketing claims to the contrary. We also expect that
one or more leading network infrastructure vendors (such as
Motorola, Nokia, Ericsson or Cisco) will also change the colour of
the market by 'moving up the stack' and delivering higher-level
middleware services.
www.ovum.com