Description
On a recent video on channel 9 i left a addmittedly rather salty comment basically noting that most if not all talks, blogs and general information about service fabric is in regards to mesh, containers or some combination there of, and that there had been very few features added to the regular app models, Reliable services and Reliable actors, for quite some time.
Microsoft pushed heavily for these app models, yet they are completley unsupported on mesh with no public road map for when that will change. We were involved quite early in the SF public lifetime and we were assured that reliable collections and actors was the future and something that microsoft would invest significantly in for a log time (hence my saltyness)
Im now of the impression is that these app models, while supported in terms of security and stability, will not be getting any significant feature updates.
Am i wrong? Have i missed a build session or blog post where the future for these app models have been laid out? One major gripe we've had for years is the backup api where you have to provide a physical path for data to be written/read to instead of a stream or some other lower level api. Another is being able to programatically access and edit backup data (for the purposes of restoring data to a diffrent number of partitions) as well as explore data in the cluster. Another one is lower level acess to replicated storage. Stand alone reliable collections are another one. Restoring to single node dev clusters is another..
These are things that the team has said that they're been working on for years, and while i have no doubt that's true, or that the team would like to deliver them, they have not been delivered.
So what is the future of the Reliable* models? Are there any planned feature updates? If not that's fine but then what are the migration paths? For our part we're talking the lack of information as a bad sign and our long term planning has us moving from reliable services initially, with reliable actors following some time after that.
Since we're so heavily invested in these app models (especially actors) however this will take alot of time and effort that we much rather spend elsewhere, but sadly, as much as i want to be wrong, we've seen this pattern before and with no information, we're left guessing.