Comments on: A world without legacy systems

By: Steveny23

Steveny23 — Fri, 30 Jul 2010 18:54:12 +0000

Robin, while I agree with your comment in practical terms, I find it rather funny Since this is a CIO forum and being one myself, i think I look at the way CIOs define legacy apps: legacy app is the one whose vendor had been acquired, the product has been decommissioned or has limited to no support options and no future upgrades with limited migration path. I do agree that a lot of the “black boxed” CRM systems for example become legacies the moment they are deployed in production. This is actually done by design and achieved through vendor lock in and most importantly by jailing your data in proprietary formats without any portability or providing a secure and universal access to the information from the outside (other apps). This is a real problem which applies to both stand alone applications and application development platforms.

By: RobinGoldsmith

RobinGoldsmith — Fri, 30 Jul 2010 18:31:21 +0000

The proper definition of a legacy system is “one that works.” As soon as the whiz bang new system you’re contrasting with some bad old “legacy” system goes into production, it becomes a legacy system too; and someone will start complaining about it immediately.

By: Steveny23

Steveny23 — Fri, 30 Jul 2010 18:05:35 +0000

Kristen, I second this completely. I spent a good portion of my previous life with big consulting firms helping customers to jump from legacy apps to new systems all the time; only to find out a year later that the “new” system had become a legacy itself. Fierce M&A activities in mid-90s and Oracle shopping spree definitely fueled this practice for years. Some consulting companies built their practices around converting legacy apps.
I think the two main issue remains to be solved before the legacy problem and vendor dependency in general can be addressed: that is data portability with universal access. Open source software never completely addressed this problem because they focused on the application themselves and not on the underlying data. The new solution which emerged recently, I believe, can actually address the vendor dependency and legacy upgrade treadmill by making access to the legacy data universal: its called data virtualization (DV). DV can actually “unleash” enterprise data from being locked in silos and application repositories, without the need to go costly and length data warehouse route or in fact having to move the legacy data anywhere. Once you solve the data dependency and data lock in problems – you are not dependent on the legacy anymore as the data becomes portable to any application. In fact, the value of the apps becomes de-emphasized while portability of the data, universal secure data access and metadata catalog creation including secure access permissions becomes the king.
Don’t mean to plug anybody in here but take a look at the pioneers in this space which will go far, like a Sunnyvale start up [A href="http://www.queplix.com"]Queplix[/A], or Composite or Denodo.