Curt Monash is a leading analyst of, and strategic advisor to, the software industry. He is also the author of the popular blogs DBMS, Text Technologies, Strategic Messaging and Network World's A World of Bytes. He joined us for a live Network World Chat where attendees asked him a wide range of questions on the software industry. He covered why social networks are not just an employee distraction but also an untapped marketing tool. He also talked about the best tools for enterprise data warehouses, corporate blogging strategies and text mining.

Moderator-Julie: Welcome everyone and welcome to our guest today, Curt Monash.

Curt_Monash: Hi everybody! Fire when ready! Anything from social networking or blog SEO strategy to hardcore database bits and bytes.

Moderator-Julie: While Curt is typing answers to questions, here's a pre-submitted one: What products are best for enterprise data warehouses?

Curt_Monash: For many enterprises, their main OLTP DBMS will do just fine. That's less likely for bigger enterprises. But the reason to have a separate data warehouse at all is performance, response time, and so on. If your database is small enough, the best answer might be to throw hardware at the problem and not split out a data warehouse at all. Or you could do what many enterprises do, and have a second instance of the OLTP DBMS, configured as a warehouse.

Beyond that, it depends on what you mean by EDW. If you mean exactly what Teradata does -- which is high volumes each of multiple kinds of analytic processing -- Teradata could be a great choice. Netezza is ever more mature and ever better at concurrency, so it can handle an EDW for many enterprises. Generally, if you're going to look up or update single rows a lot, row-based systems are better than columnar, but columnar vendors insist they do OK too. And more mature products usually handle more concurrent users than newer ones can. I have a new white paper up at http://www.monash.com/whitepapers.html. It's called "The Explosion in DBMS Choice" and it lays out when one should use something other than Oracle, SQL Server, or other marketing leading OLTP DBMS. I'll also post the intro to that white paper to my blog at http://www.networkworld.com/community/monash.

JeffCaruso: Aren't sites like Facebook and Twitter just a part of life now, work or otherwise? It seems pointless to try to block them. Would be like whac-a-mole - smack one, and another site pops up.

Curt_Monash: Absolutely. Blocking the whole site is silly. One can monitor usage to cut out the worst of work slacking, but that's about it.

JeffCaruso: How do you recommend users monitor usage and act on it?

Curt_Monash: Good question. If they have Websense or the like, that would seem to be the ticket. Websense is better for watching than for outright blocking.

Mike: Can you site some examples where Web 2.0 was used to build business?

Curt_Monash: It's built O'Reilly's conference business really well!

Mike: Can you site some examples where Facebook, MySpace and other social networking sites were used to build business? I'm just looking for some examples that I could present to management.

Curt_Monash: Mike, I'm not sure I can help you with that. Blogs are IMO much more important than Facebook/MySpace. And if the latter are useful at all, they're for faddish things like rock bands.

CuriousCat: What's the most interesting social networking tool you use and why do you like it?

Curt_Monash: WordPress. As far as I'm concerned, blogging is by far the most important aspect of social networking. It lets me communicate however I want, on whatever schedule I want. It rocks. It's pretty much the only way I market, so I can say that a huge fraction of my revenue is derived from blogging.

JeffCaruso: You blogged a few weeks ago about Microsoft's purchase of DATAllegro. Could explain why that's a big deal, and if you have any further thoughts on what we can expect to happen as a result?

Curt_Monash: Microsoft SQL Server and Oracle are simply obsolete technologies in the area of very-large-scale data warehousing. If you have 100 terabytes of data to manage, it is supremely painful to use those DBMS, if you can get it done at all. But Teradata et al. can get it done a lot better, and et al. includes DATAllegro. MPP (Massively Multiparallel Processing) is simply the way to go. More, cheaper parts is MUCH more cost effective than fewer more expensive ones. And even more to the point, disk rotation speeds and hence the hard limit on seek times have improved 12 1/2 fold since IBM introduced the Winchester Disk Drive during Dwight Eisenhower's first term as President.

CINDY: How can a company incorporate social networks into lead gen activities - has anyone done this successfully?