This post originated from an RSS feed registered with Java Buzz
by Michael Cote.
Original Post: SaaS IT Management from PeopleOverProcess.com
Feed Title: Cote's Weblog: Coding, Austin, etc.
Feed URL: https://cote.io/feed/
Feed Description: Using Java to get to the ideal state.
Of late, the topic of IT management for Software-as-a-Service (SaaS) or internet-hosted applications has come up in my conversations frequently. I took the time to jot down some notes in a mind-map:
There are few points worth calling out:
By “SaaS IT Management” I mean managing and ensuring the proper health of a company’s SaaS assets. That is, the classic trope of making sure all that “computer stuff” is running and working so that the business can do it’s think and make money.
I do not mean providing IT Management as a service, as in the case of FiveRuns, Klir, and Versiera. That topic is rad, but it’s for another bucket of thought.
The changing nature of SaaS - in my view, SaaS is the technical under-pinning of a wider shift towards using primarily hosted applications (or applications that merely have desktop clients vs. “living on” the desktop) instead of in-house applications. Culturally, there are many, many more issues involved than just monitoring and managing uptime. For example, how do local laws globally apply? If you’re getting a service for “free,” what do you even do to “manage” that?
How do you evaluate lock-in to a service? Sure, you can probably export data, but how much will it cost to import that data into a new service.
In the area of identity management, how do you do SSO across consumer and business accounts? Do you (try to) force people to use official “work” identities? What do you do if they don’t?
With mashups, composite applications, and other in-browser middle-ware how can you avoid endless finger-pointing?
Theoretically, user-behavior tracking is much easier in SaaS offerings than desktop applications. How can you use this data to increase sales or quality of service. Amazon, for example, can track how customers take to new features and changes as rated by revenue. Can SaaS consumers benefit from the same “tests”? If so, what does the SaaS IT Management software need to do?
Along those lines, part of the benefit of SaaS is more rapid and frequent updates. How does provisioning and change management work when you can have multiple updates in one month, one week, or one day?
If so many of the concerns are the same, what is so different this time that it’d be worth it to move to a SaaSier model? What assumptions do justifications make that may be hidden, e.g., Google or other service providers being able to see all your precious corporate data (vs. encrypting it on their side)?
How can you overcome IT department resistance to change to move from in-house to SaaS? Is that resistance justified because you’re going to can all/most of them?
Obviously, I’m at the questioning point ;> (I’ve also got a bit of a cold so the gears are missing a bit).
Anyone have other issues or commentary?
Disclaimer: FiveRuns is a client. As are several folks mentioned in the mind-map: IBM, parts of Microsoft, and A9, a subsidiary of Amazon.
Previous Topic
