With the launch of SQL Server 7.0 from Microsoft, the
RDBMS market promises never to be the same again. And spearheading this global onslaught
against well-entrenched heavy weights like Oracle and Informix is Jim Ewel, Director, SQL
Server Marketing, Microsoft Corp. For Ewel, who has been with Microsoft since 1989, such
formidable challenges have become a routine now. From 1989 to 1994, Ewel was a Technical
Manager in the Chicago branch office of the company. In 1994-95, he launched Microsoft's
NT Server 3.5 and BackOffice 1.0, taking on the Unix and suites world. In a long-drawn
discussion with DATAQUEST, Ewel countered many of the presumptions floated by the
heavyweights and gave an insight into the aggressive positioning of Microsoft's latest
release into the RDBMS market. Excerpts:
What prompted Microsoft to get into the relational
database market?
David Vaskavitch who is the Vice President of Development at Microsoft, went to Bill Gates
and Steve Ballmer and said-look we got to get very serious about the database market.
Steve and Bill gave him the go ahead to hire some very serious database programmers. When
we got into the Windows NT market, we hired Dave Cutler, who had a lot of experience in
building the previous versions of Unix, Digital VMS and other operating systems. In the
database market, we did something very similar. We went out and hired Hall Berenson and
Peter Byrne from Digital, Goetz Graefe, Chief Architect of Informix, Bill Baker from
Oracle and Jim Grey.
When did this take place?
This was four years ago. We began to put together a world-class development team. These
people came on a board and with SQL Server 6.5, two-and-a-half years ago, we sort of got
our feet wet, put out a version and made some changes to it. But version 7.0 is really the
first version that has Microsoft's real stamp on it. It's a version that we feel is a
breakthrough product, not just for Microsoft but also for the industry in many ways.
Microsoft brings some unique strengths to the database market that are very different, for
example from the Unix vendors.
How would you describe these particular strengths?
Our strength lies in ease of use, which has been a traditional Microsoft strength. If you
look at Oracle, their business model is based on complexity. They earn so much of their
revenue from service and from complexity of their product. We have taken an approach that
customers want ease of use both at the high end as well as at the low end. Customers in
enterprise accounts tell us that database administrators (DBA) are expensive and far too
few. They want a product that requires fewer DBAs and where routine tasks are automated
and DBAs have time to concentrate on other things. At the low end, companies typically do
not have DBAs.
According to the competition, enterprise software is not
Microsoft's core competency. So how does the launch of BackOffice Server range of products
mesh with your core competency?
Microsoft's core competency is writing great software and we have been in that
business for a very long time and have been very successful in OSs and applications.
...which has primarily been in the desktop arena?
We have been very successful with Windows NT, which now out-ships all versions of Unix
combined. It is really funny, that they
enterprise segment, because you can't find any enterprise company that does not use
Microsoft software. Microsoft with Windows NT has become very successful and now with MS
Exchange, which scales much higher than Lotus Notes, the number two player in that
marketplace. Now Microsoft has a real break-through product in the database segment.
One thing that we are seeing is that the criteria is changing and there is a move toward
integration, toward having an integrated product. This is a traditional Microsoft
strength. If you look historically, Word Perfect and Lotus were strong word processors and
spreadsheets, but they never did understand the concept of integration of suites and the
integration that it takes to build a suite. We think that there is a very strong
possibility that history will repeat itself. The database vendors, who only think in terms
of database, won't be able to come out with an integrated product.
But then they are trying...
Well, Oracle tried and then they cancelled the project, so I don't think that the company
is actually trying. IBM is trying to do that, but they have not established a
best-of-breed product. What Microsoft brings is the ability to establish
both-best-of-breed products and products that are very well integrated together. So, for
the customer it means less cost of management and actually less cost of ownership.
What is the current market position in terms of license
shipments?
Last year, we shipped four million clients of SQL Server worldwide, while Oracle shipped 6
or 6.6 million clients. And so, we are the number two database vendor in terms of units
sold and number three vendor in terms of revenue. We are the number one vendor on Windows
NT. Oracle, probably, because of its high prices is a few percentage points ahead of us in
terms of revenue, but in units we are clearly the number one vendor.
Microsoft SQL Server shipments are limited only to the
NT environment...
I am talking about six million clients of Oracle across all platforms and four million SQL
seats on NT. Oracle has actually tried to make this into a platform for discussion, but I
think customers are looking for a solution. Regardless of whether the solution runs on NT
or Unix, they are looking for the best solution and that is what we are trying to offer
them. We think that by integrating our products really well into both Windows NT and
Windows-we also offer a desktop product-we can offer them the best solution.
With Oracle well entrenched in the verticals, how do you
plan to get the application developers move to SQL Server 7.0?
Customers don't want just a database. If I gave you a database on Christmas under the
tree, then what would you do with it. People want solutions and databases require
solutions. What we are seeing are organizations moving more and more to package solutions
rather than writing their own. So our strategy is very clearly to work with application
developers. Over a-year-and-a-half ago, we founded a new customer unit. Their charter is
to work with application developers and make sure that they have the best information to
write their applications on our platform, and especially SQL Server. We have a separate
section on the Microsoft web site on industry-specific sections that gives you details
about the application and how to get them, the application developer and what products
have been used.
One of the areas that organizations are beginning to look at closely is to set up linkages
with their customers and partners. If you build automobiles, for example, you have many
suppliers. Rather than keeping large warehouses with parts of your own, you want to keep
the inventory in their warehouses. And you want to know the inventory, more or less on a
real-time basis or at least on a daily basis. You also want to provide your suppliers with
the information of how many cars you are manufacturing and how many parts they need to
supply. Thereby, automating the complete supply chain.
Are you talking about extending the supply chain?
Yes, not only is it extending the supply chain but also automating it.
This would be highly dependent on application developers
using SQL Server to build such solutions?
One of the reasons why SQL Server is such a good solution is that in any supply chain
you have both large suppliers as well as very small suppliers. Take a company like Boeing,
for example. You do not have to go very far down their supply chain before you find a
company with four to five employees. A company that small cannot handle the complexity of
Oracle-so a product like SQL Server 7.0 is better suited for their purpose.
One of our advantages, as I said earlier, is about ease of use and another one is of
scalability. Oracle always talks about this being an advantage for them, but what they
mean is that they run on extremely expensive high-end machines. What we mean by our
advantage of scalability is a range of scalability. We cover a wider range of customer
needs than Oracle does. Sure, Oracle is great for super computers, but what we offer the
customers is a wide range of scalability.
Probably we can link it to the cost of ownership, where SQL server is better than Oracle,
in fact much better.
A complete rewrite was done on SQL Server 6.5, prior to
the release of version 7.0. What about backward compatibility with SQL 6.5?
First of all, it was not a complete rewrite of the software. It is a major rewrite
upgrade of the product. Our goal with the product was absolute 100% compatibility and we
did a number of things to ensure that. One of the things we did was to ask customers for
their existing databases and their existing applications. They gave us 1,200 databases. We
converted them in our labs. About 99.8% of these were converted without any intervention.
We also have been working with major ISVs for many months to make sure their applications
work well with SQL Server 7.0. We expect over 300 applications to ship within the first 90
days and over 3,000 applications to ship within 12 months of SQL server 7.0's
availability.
Does SQL Server 7.0 have a more open product framework?
We define open as a customer choice, not as forcing customers to accept only one solution,
for example Java. Their
Java. Our solution is that if you want to write in Java, Visual Basic, VC++ or COBOL, you
can generate DCOM components using any of these languages. We think this kind of extended
customer choice is much more open than to tell programmers that they must forget
everything they have learnt and get trained in Java. It's just that all the world's
programmers do not know Java at this point and we think that multiple choice is more open
than forcing a single choice down a customers throat.
What do you think will be the main drivers for
application developers to start moving applications to SQL 7.0?
I think the wide range of scalability will be a major reason and some of the features
for business intelligence or datawarehousing market. We have taken a different approach
from the Unix vendors.
How have datawarehousing features been integrated in SQL
Server 7.0?
We include all the pieces of datawarehousing as standard parts of the product. We don't
charge extra add-ons and it is well integrated into the product as well. We have included
data transformation services in the base of the product rather than as an add-on, that is
not integrated. We also include a meta-data repository that actually tracks the meta-data,
the data about the data in the datawarehouse. We also have good integration with Microsoft
Office. We have automated most of the process by addressing the whole problem rather than
a piece of the problem as Oracle does.
So what share of the datawarehousing market is Microsoft
looking to capture?
Its hard to really tell because there are a few scientifically conducted studies about
share in the datawarehousing market. There are no really good studies that I am aware of.
We think, it is one of the fastest growing markets and believe that, in particular, it is
going to become a market that is not just the province of the world's richest companies.
In fact, datawarehousing is going to become available to a much broader audience and is
going to become a standard. Here, we have also worked with ERP vendors like Baan to make
sure they offer datawarehousing options with their ERP packages. Baan, for example, has
decided the only way, it will offer datawarehousing is with SQL Server. It will not offer
an Oracle solution for datawarehousing. It will strictly be based on SQL Server. You want
datawarehousing from Baan, you get it on SQL Server.
How have web technologies been built into the product?
We also differ from Oracle, since we do not look at the internet as just a database
problem. There is a saying, 'If all you have is a hammer, then everything looks like a
nail.' So to Oracle, everything looks like a database problem, however, we don't think of
the internet as simply a database problem. We think that web servers, transaction
capability, message queuing services are all important. We have an entire platform for
doing internet computing. We have the most successfully used platform in the internet
today, with Windows NT Server, Internet Information Server and Active Server pages and
Visual Studio.
We have done specific things in the internet database area like improved connection
pooling for SQL Server. This is important because typically you may have tens of thousands
of users. SQL server can handle those efficiently with only limited number of database
connections because of this connection pooling technology.
We have also implemented something called Dynamic Encryption. It is important to have a
secure database on the internet and one of the ways in which you integrate security is by
encrypting the data. If you encrypt it all the time, you pay a performance penalty all the
time. We allow you to turn on the encryption, say when you are passing on credit card
numbers or any other sensitive information and then turn it off during the other times to
get a performance boost.
Another argument is that SQL server does not have enough
security features and certification like other robust databases?
We are in the process of getting C2 certification from the government that will be a
part of our standard product. Oracle does not offer that as part of the standard product.
You have to buy a separate trusted Oracle product to get that kind of certification. We
will be delivering that certification in the standard product. We have greatly improved
our integration with Windows NT security in SQL Server 7.0. We think in a practical manner
that does improve security over a product like Oracle, which generally has a separate
security scheme from NT. This can run into problems where users have to remember multiple
passwords and where they are not very well integrated.
How confident are you about SQL server cutting through
the traditional fabric of Oracle-Unix?
The reception that we have received so far in Las Vegas and in my own tour around Asia
has been phenomenal. We are getting a lot more people to the launches than our competitors
get to theirs'. The reviews have been very positive. I have talked to customers and one
thing I hear them say, time and time again is that they are really looking for an
alternative. They feel that Oracle is just too expensive, that they are arrogant and that
they are going down a path that has not necessarily been about listening to customers.
As the director of SQL Server marketing, is there any
one singular concern that you have about the strategy?
We are very excited about the breakthrough potential of this product and what we want
to accomplish over the next six months is to work with customers and make the product
successful in many customer sites. It's really a matter of execution and working closely
with customers to make sure that they are successful with the product.