Microsoft has had a big year in terms of the number of product launches. We've seen a new server operating system, a new mail server and new client software. No longer marketing these technologies as discrete products, Microsoft is presenting a unified strategy to leverage the power of all three platforms. In an interview, Tony Redmond, vice president and chief technology officer for Hewlett-Packard Co.'s services unit and author of "Microsoft Exchange Server for Windows 2000: Planning, Design and Implementation," offered some observations on Exchange migrations, on Microsoft's collaboration strategy and why he believes open source won't have an impact on Exchange in the long run.
What are your clients saying about Exchange 2003? How much interest is there in embracing these new releases anytime soon?
Tony Redmond: You can take Exchange 2003 in isolation and view it as the best and most secure version to date. But it's better to view it in combination with the operating system and the new client. Microsoft has to educate people about the power of three things working together.
People will always have to be convinced that a new software version is good for them. Microsoft has not gotten past [the fact] that people have to wait for one or two service packs before something is deployed. So [customers] see three products and they think, do I have to wait for six service packs to come out?
Microsoft also has to overcome resistance in its installed base that any upgrade of Microsoft's is a pain. They need more education and they need to use customers who have been through this to stand up and be counted.
Microsoft has a huge installed base for Exchange. You are potentially looking at 100 million upgrades. That's a lot of inertia to overcome. Microsoft will also approach it with a velvet glove. But the hammer behind that is the lack of support for NT and Exchange 5.5, so that will be on people's minds.
Is there anything more that Microsoft can do to help administrators with security and patching?
Redmond: I've seen a lot of commentary suggesting they could do better. But what could they do? Are you talking about stopping development? They did that last year. To me, a large percent[age] of the problem is that systems administrators don't practice good security. They don't apply patches, keep servers up to date, close security holes, run antivirus, antispam software. Then when they get bitten in the backside, they look for someone to blame.
The best security is only achieved by administrators, third-party vendors and OS vendors working together.
Apart from all of the above, is there anything IT administrators should be doing, from an architectural standpoint, that might help matters?
Redmond: They can consolidate servers. The more servers people have, the more points of failure there are. With Exchange 2000 and 2003, we've seen the opportunity to consolidate lots of [Exchange 5.5] servers into fewer 2000 and 2003 servers.
No doubt that, with the first generation of servers, people deployed too many, largely because of fears of scalability. NT itself was created in an era when people said 'one application, one box,' and 'don't scale it too much.' Today, we have a different kind of Windows and Exchange. Those boxes on Exchange 5.5 and NT 4.0 are probably pretty old. They have an opportunity to do a hardware refresh, plus take advantage of SAN technology, and consolidate to reduce the cost of ownership.
Do you see common errors among companies that are moving to Exchange 2000 now?
Redmond: Active Directory is one big problem area. The second one is sheer lack of knowledge. People assume that if they are an Exchange 5.5 administrator, that they will be a great Exchange 2000 administrator. They are not. They must be trained and acquainted with a new operating system, and with new third-party products.
Which of your clients are moving to Exchange 2003?
Redmond: It's too soon to tell. The product hasn't been out long enough. It takes a large enterprise from three to nine months to assess a product before putting it into production. So far, though, I'm seeing mainly companies across the board [who are expressing interest].
What are your thoughts on Microsoft's revamped collaboration strategy?
Redmond: One of the interesting things we've seen is the success of SharePoint technology. I think Exchange public folders were not successful in any shape or form. Microsoft had several false starts there, especially with the APIs and workflow.
The SharePoint technology holds out more hope and promise for companies, large to small, than public folders ever had. We are delighted with SharePoint Team Services because it's enabled a great deal of peer-driven collaboration. And we are impressed with SharePoint Portal Server for its ability to make a portal react in the way we need it to react. Companies can brand [a] portal with a look and feel.
It's too early to say anything about [Office Live Communications Server]. [Microsoft is] up against some heavyweights in the market, including [open source] Jabber and IBM Lotus.
Apart from that, look at Conferencing Server in Exchange Server. It's been dropped. Microsoft is focused on Exchange being an e-mail server.
So I'd give an A+ for Exchange and e-mail. I'd give a C- for public folders and an A for SharePoint.
We've got several years before the next version of Exchange is available. Do you think this may leave some opportunity for customers to check out open source technologies for messaging?
Redmond: I don't see the open source movement stealing from Exchange. I think it will consume more of the low-end Unix installed base -- boxes running Solaris running Sendmail. They will be picked off by Linux real fast.
There isn't a lot of migration from Windows to Linux. There is a lot of hype but no one is doing it. My view is that Microsoft will retain the type of installed base it has today.
FOR MORE INFORMATION:
Conferences: Enterprise Messaging Decisions
Featured Topic: Exchange 2000 vs. 2003
Expert advice: Migrating from Exchange 5.5 to Exchange 2003
Analysis: Users get Microsoft to blink first