Backup planning for mission-critical devices


#1

Hi,

At my company we’re using Resin for a mission-critical function, a decision I inherited from predecessors. Having said that, it’s a really impressive product and I’ve enjoyed getting to know it.

But, we might genuinely be scrambling to ship product to customers if our resin devices stopped being available at short notice. Resin.io seems to still be in the VC-funded startup phase of existence, so as customers I think we have to assume that there’s still a reasonable risk that things won’t work out and Resin might not be around in the medium to long term. I’m worrying about how to communicate the potential risks to our senior management and what recommendation to make.

Have other customers here had similar thoughts, and come up with plans for how they could keep their devices running and move away from Resin at short notice? Looking at how we manage our devices right now, we seem tied to Resin for the most basic functions. And we don’t have the internal development resources (two developer for the entire stack) to feel comfortable that we could re-engineer in time to keep things running.

Finally, I don’t want to give the impressions that I’m making accusations or raising FUD. Like I said, I really like what I’ve seen of Resin so far. Just hoping for an open discussion about business risk.

Mark


#2

Hey there! It’s great to hear that Pact is still going strong. You might not know this, but Pact was the first ever production use case of resin.io, back in 2014. In fact, we hadn’t even linked to our signup page at the time, as resin.io was still in alpha, someone the team at the time found a way to it and signed up.

While we’re in our VC-funded stage, we’re also working hard to approach break-even and we take our customers’ businesses and devices very seriously. We currently have more than a hundred customers in production, and that number is growing very fast in 2018. These customers include very large Fortune 100 companies, and they would not trust us if they didn’t trust our commitment to keep the platform going for a very long time.

All that said, we’re also making architectural and licensing decisions that should ensure that even in the worst case scenario, our users and customers are not left in a rut. For one, resin.io devices are built in a way that will keep them functional even if resin.io was to suddenly vanish from existence. The ability to remotely manage and update the devices would be of course not there, but the devices themselves would keep running whatever container they had, like normal. In addition, we’re in the process of open sourcing a minimal form of the resin.io server, and of course all our on-device software has been completely open source for years. The result of these two things is that resin.io users and customers will always have the option of setting up an alternative server and reprovisioning their devices.

For customers where these above steps are not enough, we also offer an on-prem version of resin, which of course is a more expensive offering, but which does not depend on the resin team to keep running.

All in all, we feel grateful to our customers for entrusting us with their devices, and we are making every possible choice we can to assure them that both organisationally and architecturally, we’re committed to ensuring their devices keep doing what they should for a very long time to come.


#3

Thanks for the very detailed reply, Alexandros.

2014 was a long way before my time at Pact, so that’s a really interesting story that I hadn’t heard before.

With the limited development resources Pact has, I don’t feel that we could bootstrap a reliable system from the software you’ve currently open-sourced. And our Pis run in an environment where they’re being regularly unplugged and restarted - not always intentionally - so most weeks we have to manually intervene through the web dashboard and / or SSH at least once. In other words, we don’t expect our devices, as they currently work, to keep running very long after Resin went offline.

So I’m really interested to hear about the open-sourced version of the resin.io server that you mentioned. We’d definitely like to know more about that in terms of both features and time-scale for release.

Mark


#4

Hi Mark, we’ve just posted our latest update on open source. It’s not going as fast as we’d like, but we’re making progress – https://resin.io/blog/open-source-resin-io-progress-and-next-steps/