Amid a flurry of posturing and sabre-rattling, most
notably from those in the anti-Microsoft camp, Oracle has announced the launch of a new
database server called Raw Iron. The product comprises a refined version of the Oracle 8i
database running directly onto Intel or Sparc hardware, with a small operating system
kernel. It will be bundled with a hardware server, providing an easy-to-use database
server with the operating system hidden from the user. It is being seen as an attempt by
Oracle to take the database ground away from Microsoft at the high end even before it
becomes properly established there.
Raw Iron is a sound idea
Opinions on whether Raw Iron will be a success are mixed, but the general consensus is
that it is a sound idea based on proven technology meeting a definite user need from a
company with strong marketing skills, and therefore the prognosis is good. However, Oracle
has to move fast before market interest slips into scepticism. The idea of separating the
user from the complexity of the underlying technology was the force behind the original
development of Windows in the early eighties, and it is still the raison d’être
behind many new technologies. Raw Iron is no exception. However, questions arise about
which type of customer and business is likely to find Raw Iron appealing.
Oracle is said to be aiming it at medium and large enterprises, but as Ian Stephen,
Software Product Manager with Compaq points out, "Many organisations are reluctant to
have machines dedicated to one application, and also want their servers to perform other
functions and run other applications. It would suit a large organisation which wants to
have dedicated servers, but that does limit the potential market considerably." Any
firm which also wants to use the server for network administration, for example, would
have no interest in Raw Iron.
Power on the desktop or server
The question comes down the old Microsoft versus Sun debate of whether the power should be
on the desktop or at the server in a thin client model, and it is hardly any surprise that
the first vendor to sign up to the Raw Iron project was Sun. Other vendors are already
showing interest, and Compaq’s Stephen says, "For us it will just be another
operating system and application running on our boxes. We don’t see it as such a big
deal, although it is an interesting idea and one with plenty of potential." Digital
pioneered the idea of a dedicated database box in the 70s but the market demand for
time-share computing with more flexibility in each box ensured that the products never
really succeeded. However, Compaq’s Stephen believes that IT mangers will be more
receptive to the concept now. "The propaganda about network computers has cleared the
way for dedicated servers," he says.
Other pundits, like David Wollard, director of technology at Valtech, a consultancy
specialising in solutions based around object technology, says, "Oracle has simply
realised that users don’t want to have to deal with an operating system, and a
visible operating system is unnecessary in most instances. The Oracle database provides
all the necessary functions and the operating system can be stripped down to the bare
kernel, and the whole thing can be packaged in a single box which is simple for the user
to get to grips with."
SQL 7 or Raw Iron
Wollard sees Raw Iron as a robust and extremely credible response to Microsoft’s SQL
7.0, "in terms of reliability, scalability and stability, in fact all the things that
today’s IS managers are concerned about." But still Oracle is going to find it
an uphill struggle to usurp Microsoft in the market for medium and large organisations.
"The problem for Oracle is that Microsoft is brilliant at marketing, and already
there is a tide of confidence that Microsoft is going to be able to deliver robust and
scalable products. But Oracle is also very good at marketing, and has the advantage that
it is attacking Microsoft at its weak point, at the top end of the market and at the
database and operating system level, so in my view, the jury is out. We’ll have to
wait and see which punches harder and which products survive the reviews and tests
best." Wollard is extremely positive about Raw Iron, not least because it removes the
importance of the choice of OS provider from high on the list of strategic decisions. He
adds, "Making Solaris the first platform to bundle with Raw Iron is a great decision,
not least because Solaris is so well established and so proven and potential customers
will feel confident and safe with it."
Terry Lawlor, UK Marketing Manager for one of Oracle’s competitors, database provider
Informix, takes a predictably cynical view of Oracle’s ability to deliver the product
it has described and promised, but still finds it hard to criticise the basic premise of
the technology. "I believe that there will come a time when database technology
services are packaged with cut down operating systems and shipped in high volumes, but I
think that time is still a long way off." Lawlor also points out that there is
nothing fundamentally new in the philosophy behind Raw Iron but he says it is something
that resellers and solution providers can deliver. "Users are looking for a turnkey
system that they don’t have to think about, and those are already available," he
says.
Doomed to failure
One IT manager at the coal face, Dominic Connor, head of IT with City-based bond brokers
King and Shaxson, thinks that Raw Iron is a great idea yet still doomed to failure. He
explains, "Raw Iron is technically a great idea. I manage a pile of servers which
have piles of functionality which I will never use, and the fact that mission critical NT
servers also support user-functionality worries me a lot." Connor believes that PR
hardware is weak at managing databases, and points out that several manufacturers like IBM
and ICL have already managed to put enough intelligence into their disk controllers and
read heads to allow searching to be done by the disk unit itself. "It is also
difficult and fiddly to make databases secure, which is a big plus point for Raw
Iron." Connor explains that in his organisation all passwords are changed once a
month, and "users must be coerced into changing their NT, Novell and Sybase
passwords. If they set them differently, they lose one or all of their services, and
automating password changes is dreadfully hard and unreliable. In Raw Iron there would
only be the Oracle log-in, thereby removing this problem."
Yet Connor says that despite its ease of use, and the big bonus of inherent security, Raw
Iron doesn’t stand a chance. "The average user just wants more than the Raw Iron
box will offer. It’s all very well cutting things down, but users want to have
support for NetWare, for example, and backup software, and drivers for the network cards,
RAID controllers and SCSI cards that the system will have to connect with, not to mention
support for SNMP."
User scepticism
One of Ciga’s analysts, Mike Sun, says that Raw Iron looks good on paper,
particularly the fact that it will reduce the cost of IT ownership in several ways, but:
"they are going to be caught in a classic Catch-22 of user scepticism until they can
get customers to stand up as reference sites and talk about evidence of cost savings and
show proof of performance gains." Consequently, Sun believes that Raw Iron will
struggle to gain a toehold of sales. "Customers are tremendously cautious these days,
and although all the benefits look good there could be lots of drawbacks that aren’t
even evident yet, such as the inability to use third-party tools and the risk of creating
even more legacy data and systems."
And Sun agrees with Compaq’s Stephen that many users will not want a dedicated
machine that they can’t use for third-party tools and non-database applications.
"It’s a matter of flexibility of IT investment, and buying a machine which can
only be used for one purpose is going against current trends and thinking," he says.
But he adds, "Customers are also looking for lower computing costs, and with an
operating system that is geared to do little more than be an interface between the
database engine and the actual hardware, one expects that there is less to break down and
there should be better database performance." Sun also expects Raw Iron to be able to
deal with any type of file, including spreadsheets, presentations and emails, so that the
argument that a Raw Iron box is too application-restrictive loses its force.
Dedicated database servers are an attractive proposition
Other pundits believe that there are plenty of customers who would be keen to have a
dedicated database server. Kevin Adams, managing director of telecoms software specialist
Intake Systems says, "Dedicated database servers are an attractive proposition for
companies that want absolute maximum database performance, particularly in three-tier
client/server architecture." He believes that it will be good for systems which run
OLAP or for data warehousing applications, "where there are high numbers of SQL
queries to service and manage." Adams also expects Raw Iron to be particularly
appealing in customer care and service applications. "Call centres are one of the
fastest growing areas of IT and large numbers of users need to be supported easily and
with extremely robust and fast response times," he says. "In our business too,
telecom billing systems, large quantities of data have to be stored and accessed, and
scalability is vital. A dedicated server is necessary to deliver the reliability and
confidence in the system that we need." However, Adams doubts whether Raw Iron will
make a significant dent on Microsoft’s march from the desktop to the enterprise.
"Even if Raw Iron is successful, I’m not sure that it will hurt Microsoft very
much. After all, someone has to supply the desktop clients to soak up all that raw SQL
power."
Microsoft isn’t worried
The last word comes from Mary Hope, a consultant with independent analysts and researchers
Ovum, who believes that the cake is big enough to support all the technology options and
all the vendor players. "I’m not sure that Oracle really cares whether everyone
continues to buy NT and Unix, because Raw Iron will eventually run on those as well as
Solaris and Sun. And I’m not sure that Microsoft is going to be too phased if Oracle
has success at the medium and large corporate level, because its position is secure and
growing in those sectors regardless of what other technologies and vendors are
doing." Despite reservations that Raw Iron is not as innovative as Oracle is
claiming, she believes that the philosophy and technology are sound and will appeal to a
large enough base of potential customers to justify the product. "I’d be
concerned that Oracle is trying to reduce the architectural options available, and I
wonder if that is always good for the customer," she says. "But on the other
hand I doubt whether Microsoft is feeling nervous about the launch of Raw Iron."
Valtech’s Wollard agrees that there is still much unknown about Raw Iron to make an
accurate judgement. "There are still too many vital questions that we don’t know
the answers to. I’m not convinced that enough businesses with diverse needs and
functions will want a single-function server. And there are questions about how Raw Iron
will get to market because resellers won’t be attracted by a new machine which
doesn’t offer much margin." But Oracle’s reputation at the high end as an
enterprise level solution provider, just the area where Microsoft is focusing its target
these days, may be enough to have Microsoft rattled. Oracle has a healthy obsession with
maximising product performance and reducing cost and complexity, which IT managers
generally find appealing. Ellison has made no secret of the fact that he is targeting Raw
Iron directly at Microsoft’s jugular, as he sees it, and the attack comes at a time
when the industry leader is particularly vulnerable. "We will compete with a lower
total cost of ownership, lower purchase price and higher service, faster application speed
and greater reliability in the application environment," says Ellison. "This is
not a marketing stunt. This is a serious effort to reduce computing costs by simplifying
systems management."
Wait until next year
At the moment we have yet to see Raw Iron, and it is still unclear how many hardware
vendors will have signed up by this time next year. For Raw Iron to be a success, that
list has to include a healthy majority of PC vendors, and the consensus among them is that
Raw Iron is a good idea, but involves marketing and product complexity that they can live
without. It is possible that Raw Iron will fall at the first hurdle and not even reach the
market, going down in history as another piece of rhetoric from Larry Ellison in his
personal feud with Gates. One thing is certain - Oracle needs to get some customers
interested fast and then use them as reference sites, before the rest of the world will
even take the idea of Raw Iron seriously. |
|
