Microsoft embeds instant messaging into OS

The sands are shifting again at Microsoft as highlighted by its Greenwich project that pulls instant messaging support from Exchange and transfers it to the base operating system. Also, Microsoft just announced its Jupiter project to consolidate three .Net business servers , and XSO an API for accessing Exchange features. Paul Flessner, senior vice president of .Net enterprise servers at Microsoft, sat down with Network World senior editor John Fontana during the Microsoft Exchange Conference this week to discuss how and why all these changes are going on.

You said in your keynote address that Greenwich is a statement of direction, can you explain that a bit more?

Sure. We made a hard decision probably two years ago to pull [instant messaging] out of Exchange, but we are going to honor all of our commitments to our Exchange 2000 customers. Those customers will continue to get the same functionality they have had moving forward. But new Titanium customers that did not have Exchange 2000 won’t get instant messaging or RTC [real-time communication], that will come with the platform. The reason we made the shift is we realized it was going to be a big wave and it’s really a piece of platform technology that we think all of our customers should have not just Exchange customers.

Define “big wave”.

Thirty-percent of business desktops use some sort of unsecured form of IM today. And I do think video conferencing, presence and location and all that stuff, start to look very technically like traditional networking, so we decided to pull it away from the Exchange guys and put it with the Windows networking team. It was the right decision technically. We do have some business model stuff to work through. IM is a big play. You couple it with SQL Server notification services and when you are doing publish and subscribe and pushing things out with location services and presence services and you start to tailor it to the device – I think it is a killer.

But this switch also allows you to sell IM without having to sell Exchange. So if I’m a Domino customer on Windows you are not going to lose out on IM to Lotus by default?

If you are an Exchange business guy you wanted IM to stay because it was a pretty big value add. If you are Microsoft and you want that kind of technology to be propagated broadly, like the guy at my last meeting said ‘phones only work when you have two, one phone doesn’t work’ so we have to have it as a technology that is propagated very broadly for it to be successful and that was the choice.

Last year in your conference keynote, you said part of your talk was owed to the Exchange guys because you were changing their backend storage, do you feel you owe them another talk about IM?

It’s mostly plumbing and I think Exchange will add value on top of it. So I don’t think it is a big takeaway. And Realtime, whatever the one is from IBM around Notes [ed note: Lotus Sametime], clearly they still sell it that way but clearly they are not a platform company and they don’t have the same pressures we do around the operating system. But I think it was the right thing to do, I feel good about the decision.

When you said you would take care of Exchange 2000 customers, will you take care of them when they step up to Titanium?

Yes, they will be taken care of if they were an Exchange 2000 customer. They will get equal functionality going forward.

Even if they go to Titanium but chose to run it on Windows 2000. In that scenario you have migrated yourself out of instant messaging?

No. IM will never be taken away from current Exchange 2000 users, I wouldn’t even think of it. I will take care of it myself even if I have to fly in and deliver it to them they will have it.

Is there a de-emphasis on Exchange. You added the Web storage system, IM, conferencing and they are all going away now?

No. We’re absolutely investing like crazy in Exchange. This [real-time communications] stuff has just been put where it should be. Look, Exchange and Notes all grew up in these stove pipe environments, they had their own store, their own protocols, their own development tools and, you know what, they got to market maturity. How do I grow from a 61% market share in the enterprise, it is tough. What am I going to do, make a better proprietary tool. It’s not interesting. Your only choice is to face it, which we did two years ago, and we were very open that we were going to re-platform the thing on top of SQL Server, that we were going to make it a first-class Web services system with the tools to support it. Is that a threat to the Exchange business? I don’t want to insult you, but you have to think more broadly about what mail can be. It is a dial tone. It has to be treated like a dial tone. More integration with messaging and with the client, you bet. There are a lot of things I can do there. Integrated voice mail, yeah, probably, that is kind of a cool idea.

So the platform breaks down to that core messaging, the focus is making it solid and all these other pieces go away?

I used to have 600 people on Exchange and now I have 350 and we are going to have a product that will do more rather than less. You know why? Because I don’t have to build all that platform stuff anymore. They don’t have to be protocol guys, they don’t have to be storage guys, they don’t have to be directory service guys anymore, all that stuff has been doled out to guys who do that for a living. And these Exchange guys can focus on adding value to the messaging system. Certainly it will become a more reliable mail system.

When you talked about complexity in the .Net platform in your keynote you had a chart with 13 Microsoft server boxes piled on the side opposite the single mainframe box and I thought, now how does that reduce complexity?

It increases complexity, there is no question. The world is more complicated. The question is what can we do to decrease the complexity of implementing that for the customer and some of that is Web services and some of it is integrating products. The world is more complicated, there are more things to do, and they are more complicated and customers are savvier and more demanding. So it is up to us to figure out more ways to reduce that complexity by engineering the software on the front end to make it more integrated because we are not a services company, to pick on IBM a little because that is their strategy, and to do that in a lower TCO [total cost of ownership] way.

The Jupiter project and the XSO API in Titanium are those precursors to Web services functionality?

XSO is a little different than Jupiter. XSO is just a way to get managed code access into Exchange.

But the key is that you build applications on the operating system and link to Exchange using XSO?

Yup. And Exchange is just a service serving up data.

Today you do that with an API, down the road you are going to do it with Web services, but it remain the same general concept?

Absolutely.

So is the challenge now to move from the API interface to the Web services interface and how do you do it?

It is a challenge, but it is not that hard. It is just code, it’s not the hardest code in the world to write. The more interesting challenge is around Jupiter, how do we partition the features, which features are customers going to want. They are clear they want this as one product, so from our perspective this is one product to sell not three. So there are some advantages there in terms of cost. But the most exciting thing is that customers won’t have to integrate our integration server. It will all be integrated, and believe it or not, when you want to reach through the firewall and touch a customer with Commerce Server you also want, at the same time probably, to do some sort of workflow on the backend and connect to some kind of legacy system. We did a lot of testing with customers and analysts and they loved it.

One last thing, you said in the keynote, “Web services, we have no back up plan.” You made no bones about the direction?

None.

So it’s pedal to the metal?

In everything we do. I don’t know how to say it more directly. It is exactly what we are doing. I’m poking a little at IBM when I say that because they do have a backup plan. They are still putzing around. They stand for nothing except what they can get the customer to ask for. J2EE, sure we’ll do that. Web services, sure we’ll do that. If it doesn’t work, we’ll get [Global  Services] in there to fix it. It’s kind of a backhanded slap at them, but I am being truthful, there is no backup plan for us. It is a decision we made, it’s the right decision. It was a momentous change for Microsoft. We made a few bucks off the Win32 API. It wasn’t easy to get Bill’s head around the fact that the API is not where it is happening, it is the protocol. The game is over, distributed name service won. We made the decision and we are going on.