I think that proprietary applications will continue to thrive because operating software is expensive whether it’s open source or proprietary. For many folks, the difference in “free” vs “paid license” makes it not worth switching from proprietary applications to open source applications.

Another thing to consider is that hosting applications and data has gotten very cheap. For example, 3 years ago, our company was paying $1500/month for a single server, software, physical security, network security, and availability at a high-quality data center. If I were to do it today, I would just use Amazon EC2 to host the application at a starting price of about $70/month, with the same security & reliability. As the application processing and data grows, Amazon’s solution scales up and out very nicely, and the computing costs grow proportionally to traffic, processing, and storage.

However, even though the hosting is cheap, service operations, customer support, software maintenance and evolution etc have not gotten any cheaper. Those costs are starting to drive the overall costs more these days.

The big issue coming down the pipe is data portability and data security for applications hosted by 3rd parties. Once the big hosting companies convince enterprises that they can host data more securely than the enterprises, SaaS will become more and more dominant. Data portability has always been a big problem, and often represents the last available form of vendor lock-in.

I don’t think that Microsoft needs to buy Yahoo. Microsoft has the skills in operating SaaS applications such as hotmail, search, and XBox Live. They have deep domain expertise in the software that would make great SaaS applications, and they have the trust of enterprises that pay for that software. What Microsoft does not have is an executive culture that enables them to successfully run a profitable P&L unit that provides SaaS. What they should do is hive off a fully owned subsidiary to run Live as a separate brand from Microsoft, and then staff that subsidiary with capable executive managers from the likes of Google, Yahoo, Amazon, or eBay. They could buy up a few small SaaS players such as Zoho, and then integrate them into Live. The nice perceived advantage with Live is that if they ever claim that their software interworks with Microsoft licensed applications and services, people are more likely to believe them because they’re owned by Microsoft. The other big habit they have to break is charging every user a license for every installed piece of software. Live as a separate subsidiary will enable them to use other revenue models to achieve business success.

If not, why don’t SaaS folks etc. post out their source?

More here.

As the cost of actually doing this goes down, surely the source code is exactly the crown jewels that I’ll need to hang onto?

Maybe we’ll see an “open data” equivalent of open source that commits to making data, metadata, and relations openly retrievable to its owner in a structured, standardized manner. And maybe an open source license can enforce that kind of behavior in apptone operators.

http://www.krishworld.com/blog/open-source/the-need-of-the-hour-for-web-services-startups-data-portability-and-open-source/

http://www.krishworld.com/blog/open-source/opensource-can-make-your-web-app-live-forever/

http://www.krishworld.com/blog/open-source/how-to-select-a-hosted-wiki/

http://www.krishworld.com/blog/open-source/future-of-open-source-in-the-saas-world/

But I think that you need to distinguish between applications that are heavily user-oriented such as spreadsheets, text, CRM, calendars, email and those applications that operate a business but whose value is largely in the processing algorithms and data and require highly specialized skills to operate. A banking system, for example, that manages large scale transactions is not a good candidate for SaaS computing.

Maybe this is obvious, but as an employee of a software company that makes this kind of database-intensive, minimal-user system, I keep wondering whether the license-and-install model will eventually move to SaaS. While we see some small customers moving in the hosted direction, very few are comfortable or willing to operate their company on someone else’s server farm.