Where’s Jini?

It’s been a while since Sun Microsystems launched its technology
for creating software to make devices work together as a “community”. The
kick-off for Jini was sufficient to put it at the heart of any discussion of
the future of network software. It had its share of questionable concerns,
but these were met with the assurance that time would provide solutions.
Three years and counting, though, the question in the mind of developers
today is: “Whatever happened to Jini?”

Rolled out by no less a figure than Sun co-founder (and famous futurist)
Bill Joy, Jini was described in 1999 as follows: “Built on top of a Java
software infrastructure, Jini technology enables all types of digital
devices to work together in a community put together without extensive
planning, installation, or human intervention. Each device provides services
that other devices in the community may use. These devices provide their own
user or programmatic interfaces, which ensures reliability and
compatibility.”

It was quite a vision of things to come. So with the benefit of hindsight,
it seems fair to ask how it’s panned out so far.

I spoke first with Sun’s spokesperson for the technology, who acknowledged
that Jini has not lived up to the early hype that surrounded it but urged
patience and perseverance on the part of developers.

“When Jini first came out, it was positioned as device-centric,” said Franc
Romano, Jini group marketing manager for Sun. “Now, we’re trying for a
more-balanced approach toward software services. This is the major change in
positioning since Jini was launched.”

Today, Sun uses slightly different language in its official description:
“Jini technology is an open architecture that enables developers to create
network-centric services, in hardware or software, that are highly adaptive
to change. It can be used to build adaptive networks that are scalable,
evolvable, and flexible as typically required in dynamic computing
environments.”

Romano told me his company sees the software world as being in a
protocol-intensive period now, citing XML, SOAP, and UDDI as examples; but
they see the next cycle as being “protocol-agnostic,” offering an
opportunity for Jini to move into session and directory spaces now held by
the XML-related protocols.

“UDDI gives way to Jini to create a network of ’embedded things’, allowing
expansion into automobiles and the home,” said Romano.

Right now, though, his group sees themselves as still in an “early adopter
phase” in which they get Jini “into the hands” of developers. He estimated
that as many as 80,000 developers currently work with Jini, building
infrastructure, components, and services for networks, primarily in the
telecom, financial, and health-care fields.

“We still have a long way to go before Jini is mainstream, however,” Romano
admitted.

To combat the pitfalls of over-anticipation — hype — the Jini group has
adopted a market-driven approach, he said. “We’re being very careful not to
get out in front of our developers… to support whatever direction they’re
going.”

With a pending upgrade to Jini 2 planned for next year, Sun may be able to
buy some breathing room for the technology’s eventual success. But will
developers, notorious for their fierce pragmatism, exercise the patience and
perseverance to wait for it to finally gain momentum?

Word on the Street

To gain some balanced perspective on where Jini is now and could be in the
future, I asked around to see what insiders had to say about its status in
the industry.

For starters, I spoke with a pair of developers coming at Jini from
different angles. One using it for hardware, one for software.

Danh Le Ngoc is the co-founder of aJile Systems, maker of a direct-execution
processor for embedded Java applications on devices. They recently partnered
with PsiNaptic to bring Jini functionality to their aJ-100 offering.

He said his firm has received “strong interest from existing customers as
well as new ones, who have waited for a Jini solution for Java-based mobile
devices and Internet-based industrial gateways and sensors.” But he
identified a key impediment currently hindering Sun’s aproach for
small-footprint products — its reliance on Remote Method Invocation.

“RMI-based Jini is too big for small, deeply embedded devices,” said Ngoc.
“Fortunately, [PsiNaptic’s] JMatos on aJile processors has resolved this
technical issue.”

Asked where he saw Jini going in the future, he was nevertheless very
upbeat. “Jini-enabled devices powered by direct-execution processors will be
pervasive in the next two years. They will be deployed broadly at home,
factory, and enterprise.”

Aidan Mark Humphreys is the system architect for Procoma GmbH, which markets
a new Jini-based XML document-presentation tool called Chameleon, recently
adopted by Germany’s Commerzbank AG.

His take on Jini today? “It allows the programmer to think about developing
systems that expect and deal with outage, mobility, the addition and removal
of services — in short, Jini’s model of computing actually resembles the
highly spontaneous and unpredictable world most of us live in and conduct
business within. …Bottom line: Jini is good enough to create real software
with it now. And it’s getting better.”

Humphreys said he thought Sun had, indeed, made a mistake in taking so long
to pitch Jini to application developers, as opposed to the device community.

“Jini actually represents a conflict for Sun, it doesn’t yet help their
hardware sales, but it could if marketed as a software services
technology impact their market leading J2EE initiative and the Web services
area they are penetrating. I suspect Sun are looking coldly at Jini and
wondering where it fits into the picture.”

His outlook for Jini’s future is that it will gain wider acceptance but only
if managed aggressively. He pointed to Jini’s underlying JavaSpaces
framework, modeled after David Galernter’s famous tuple-spaces architecture,
as promising.

“There is already growing interest in tuple-spaces from non-Sun sources. The
Python and Ruby communities, for example, have their own tuple-space
implementation that could be made interoperable.”

Beyond that Humphreys was cautious. “Unless Sun really get behind Jini to
the extent that they did for J2EE and drum up support both amongst device
manufacturers and software developers, there is a real risk Jini, as a pure
Java technology could fade away.”

Analyze This

A pair of analysts were even more concerned about Jini’s drift. One was
hesitant, the other downright pessimistic.

Forrester Research infrastructure analyst Laura Koetzle said: “Sun initially
pitched Jini as an ideal P2P framework for small, mobile devices, but Jini
requires every peer to run a Java Virtual Machine. Small, mobile devices are
often resource-poor, which makes running a full JVM difficult and creates
substantial opportunity cost.”

She sees Sun’s own Jxta (invented later by Bill Joy and Crew, ironically) as
Jini’s main competitor. “By dropping Jini’s VM requirements, Jxta… lowers
the barrier to P2P networking entry. By remaining OS- and VM-independent,
Jxta stays flexible enough to provide infrastructure for the P2P apps of the
future,” she noted.

Distributed computing guru JP Morgenthal, author of Enterprise Applications
Integration with XML and Java, gives Sun a thumb’s down for what it has
accomplished with Jini. He states flatly, “Jini has not become one of the
significant technologies in the Java family.”

Morgenthal sees Jini as competing with many open standards movements at the
same time, such as Jxta and Web Services. “They pushed Jini hard at hardware
manufacturers to include as embedded components,” he said. “This requirement
results in additional expensive hardware and a reliance upon a technology
that has not been broadly adopted by a large user base.”

Jini in a Bottle

The testimony puts Jini in a precarious position these days. It could
recover from its stumble out of the blocks or end up pulling out of the
marathon, well out of the race. A harsh economic environment can mean gloom
for technologies that do not quickly fend for themselves, but it can also
strengthen the halest. We still don’t know how Sun will manage Jini to
maturity, but it clearly has a lot of parenting to do.

In the meanwhile, Procoma’s Humphreys should get the last
word: “Jini’s approach will live on, because it’s a paradigm that matches the
pattern of future computing much better than today’s big-selling enterprise
approaches. So even if Jini fades, I have little doubt that in five years something similar will have taken its place and entered the mainstream — but maybe based on .NET. That is the danger Sun are running if they walk away from Jini now.”

Resources

• Jini Network Technology
• The Community Resource for Jini Technology
• A first look at Jini lookup and discovery protocols
• Jini in reality
• JavaOne 4.2: Mr. Joy’s wired on the future
• A Jini voyager: an interview with David Norris of ObjectSpace
• A Jini pioneer: an interview with Freeman Jackson
• First wishes made of Jini
• Jini released from its magic lamp

About the Author

Kieron Murphy is the editorial manager of EarthWeb.