Dissecting the motivations, dangers, and potential rewards of a software program “coronary heart transplant”
I used to be sketching out architectural suggestions for a consumer venture that, after a profitable POC, was prepared for the prime-time of manufacturing software program. I finished to ask myself why the brand new arch diagram I used to be creating appeared vastly completely different from the structure of the present idea software program; even when my solutions appeared stable (they included particular measures of reliability, velocity, scale, and safety), I couldn’t shake the sensation that this felt acquainted in a foul approach. I counted the variety of replatforms I’ve both been instantly liable for (i.e. a number of Fashionable Information Stack conversions) or not directly impacted by (Information Warehouse reconciliations between the “previous” and “new” techniques) over the past fifteen years. I got here up with 13 replatforms. Of these 13, solely 4 ultimately yielded optimistic product or engineering outcomes; that could be a roughly 70% failure price amongst replatforms I’ve personally witnessed. It’s no surprise that something resembling a re-write, re-architecture, or re-thinking provides me instant pause.
Replatforms are resource-intense, time-consuming tasks that may make or break govt careers, mark the hockey stick inflection level for a company’s development, or go away…