| Developers conferences can be useful if you get to talk to the right people. I was given a pass to Microsoft Build one year because we were looking to build stuff on Azure but weren't sure which services were mature and which were not. I talked to almost every single PM who had a booth there (and at MS, PMs are also developers). I learned that if you push MS PMs hard enough and ask the right questions, most will drop the marketing facade and give you the insider's view. (after all, developers -- by personality trait -- generally hate two-faced marketing talk and would genuinely rather talk about the tech) This unfiltered insider's view is decidedly quite different from Microsoft's enterprise marketing's messaging. Attending Microsoft Build and talking to PMs helped us avoid investing our efforts in Azure services that turned out to be dead-ends. (many of Azure's GA stuff are feature complete but not truly production-ready) Short of running POCs, there exist few other low-effort means of procuring this intelligence other than by talking to (honest) Azure consultants at Meetups who have to deal with this stuff in daily production. My conclusion from the conference (corroborated with my own dev experience) was that the parts of Azure that were built on pre-existing Microsoft technology (like VMs and SQL Servers) were generally solid and could be relied upon. Whereas many new-fangled PaaS/SaaS cloud-only offerings tend not to be as battle-tested and would often fail on corner cases, so one would be prudent to think twice about putting mission-critical workloads on them. Also, one learns that despite the glossy marketing material, some Azure offerings turned to not have had any dev activity on them due to low uptake. There are still maturity issues in Azure today, and my gut feel is that most enterprises that do run on Azure mostly use their IaaS (VMs, SQL) offerings -- these are the most mature -- rather than their PaaS and SaaS offerings. The common refrain from marketing folks is that cloud development is a moving target, and what was true a week ago might not be true now (a trivially true statement but of no practical use). |