This post was written by Matthew Skelton co-author of Team Guide to Software Operability.
We recently contributed to the ‘DevOps Perspectives 3’ report from CA: Microservices and data consistency. Here is an extract (you can download the full report from CA).
The hidden value of a monolith
In the rush towards microservices, some teams seem to be unaware of the value previously provided by their older monolithic architecture, particularly if a central relational database was used. With a monolithic architecture and especially with a central relational database, conflicts arising from competing product budget streams are resolved at compile time, data load time, or integration test time: at any rate, almost certainly before the software reaches Production.
However, in a microservices world, the coordinating effect of the monolith or central relational database are typically lost, potentially allowing changes that conflict at a fundamental business level to reach Production before the conflicts are detected.
Data strategy for microservices
Unless different product teams are using some kind of shared bus or event store for cross-service consistency, we predict that the lack of a monolith or central database could result in teams duplicating data and logical entities across different data silos.
The issue of pan-organizational data consistency and data integrity is something that cannot be ignored if business outcomes are to be met and sustained over the coming years, even as organizations adopt patterns such as microservices to help achieve more rapid and frequent software changes.
To read the full article, download the full CA DevOps Perspectives 3 report from:
4 thoughts on “Implications for data strategy with microservices”
Reblogged this on MTE Advisors.
Well said – monolith, legacy or spaghetti, if it works, it provides value. Also, when everybody (and their dog) starts writing µservices things definitely won’t be only pretty and you do lose coordination.
Have you heard of GeeCON µservices conference? It happens in Sopot, this September. Since you are obviously not on koolaid, I think this event will be to your liking – we looked for speakers who did not just want to praise the µservices, but also tell about their dark sides.
Conference website: http://2015.microservices.geecon.org/ (just scroll down for speakers, programme, etc.)
Why I believe it’s interesting event: https://lafkblogs.wordpress.com/2015/08/05/reasons-why-i-believe-geecon-microservices-will-be-good/
I’d like to invite you to come. If you feel it’s a good event, worth your while, do come. If you think it’s great idea and would like to sponsor something, I’d be grateful. If you believe it is lacking somewhere, let me know. Oh, and let me tempt you with the beach – it’s 200m from the conference venue and breaks between lectures ain’t short! 😉
Thank you for your contribution to the report and it’s link as well.
Thanks, Tomas – GeeCon looks interesting. I will not be able to attend this year, but I’ll look out for the 2016 event. Hope it goes well 🙂
Thank you! 🙂 Will try to pen review after the event! 🙂
LikeLiked by 1 person