Coté Memo #075: Dealing with legacy code in a cloud native world

Get your lurn on – this weekend!

Do you want to bone up on your product management skills? Check out this two day workshop from Craftman PM. I used to work with Prabhakar and he’s anything but boring when it comes to opinions around product. Check out more details, and if you use the code COTE when registering, you’ll get $250 off!


W Austin

As I mentioned last time, we stayed at the W in downtown Austin last week. It was nice! The “wet” (W-speak for “pool”) was nice with in-pool bar service.

Tech & Work World

Dealing with legacy code


I’ve been working on a series of blog posts (which we’ll then slap into a PDF – hey, presto! Whitepaper!) on “the cloud journey.” Here’s an excerpt from the first part summarizing the challenges of “the legacy journey”:

These groups have a full portfolio of existing IT and applications that they must maintain and grow. There are many “obligations” owed to the past and they often operate under many more constraints than the other two types of teams. Their challenge with Pivotal Cloud Foundry is planning out how to methodically “slice off” parts of their existing applications and re-platform them as cloud native applications. These teams are metaphorically tasked with rebuilding the jet engine mid-flight.

Legacy teams are often looking for fixes to lingering, systematic problems they have (the relational database can no longer scale) and the effects of too much technical debt (“our system is so burdened and fragile that it takes weeks to do a release”). The challenge these teams have is that all of their time is taken up simply keeping their applications up and running, leaving them little time to work on new functionality in their application. Worse, when there is time to add in new functionality, the legacy system is so ponderous (and often poorly understood) that changing it takes much longer than it should.

To me, the challenges here are about balancing risk perfectly, knowing when to keep doing “the wrong thing” despite the allure of “the new thing.” Eventually, these teams have to choose either to “give up” or “go for it”: If the risks of making changes are too high, they must quarantine the applications in questions. Or, if the risks seem acceptable, the teams have to start systematically re-platforming and re-writing the backing services and applications themselves.

DevOps at Solarwinds

I was on a panel for Solarwinds online conference, thwackCamp. It was fun, actually. Check out the recording and a piece written about it.

Quick Hits

Fun & IRL

Movie Time

I say Amy) last week. I wasn’t expecting to like it much (we just wanted to see a movie, and Kim wanted to), but it was actually good. You know, tragic and such, but good.