Bought an imp for a little telemetry experimentation. I work in IT dev and wanted to experiment for my customers.
Reading through the forum I’m beginning to get a bit concerned that the hardware I’ve bought is entirely dependent on the existence of electric imp’s cloud (for dev and functionality).
Am I right in thinking
- There’s no way to load bytecode into the imp without going through the imp cloud (however unrecommended)
- All the imp devices in the world will cease to function (at least in any useful way) if electric imp (the company) goes under
- There is no way to set up your own cloud/local server
On point 3 I’m happy to write and run that from the ground up on my systems if the interface is available. Is anyone writing and running code independently of the imp cloud (details please)?
I’d happily buy dozens of these for my customers if I knew I (potentially) wasn’t going to end up explaining why everything has ceased to work and I can’t provide a backup connection.
100% correct. There are some opensource alternatives that solve that problem, but i found them to be rather buggy compared to the Imp.
Thank god tech companies never fail.
Any electric imp employee know of a way to cover this eventuality (however “insecure”). If you make yourself a link in the chain doesn’t that imply liability for its continued existence?
Anything electronic, especially wireless, has a risk of failure. One massive solar flare will wipe-out everything for a period of time (hours or days). Imagine if GPS satellites are disabled for two or three days. How would your own server/cloud be 100% reliable? You say that you could provide a backup connection, but will your devices reconnect and function without any intervention? Everything has a risk.
As @remcohn says, absolutely correct. Electric imp is first and foremost a service, not a hardware provider - which is rather borne out by the fact that the latest imp module, imp003, isn’t made or sold by us.
There are no backdoors which would allow you to substitute another server, I’m afraid.
We are, however, well funded (~$23m as of our series B earlier this year), have paying customers, and are shipping a lot of imps. Some very large commercial customers are happy enough with our position that they have used the imp in their commercial products.
mlseim: I don’t think that’s a valid comparison (solar flare). Also I’m not sure how losing GPS would affect an intranet. Thousands of tech companies have come and gone but we’ve never had the world’s electronic infrastructure permanently removed. I don’t think my customers would expect me to cater for the end of civilization but they would expect a backup plan if their ISP went offline for a few hours or days.
Hugo: I can see why you’ve marketed this as a service (revenue stream) but it writes it off as an option for my company. I would assume your very large commercial customers have asked for a fallback plan if the imp cloud goes offline. As a (potential) little customer I can’t ask for that. Also the cost of imp cloud access is a big unknown for users.
I’ll keep my eye open for any electric imp you may market in the future that isn’t cloud dependent (sell it for more money).
Excellent tech., dangerous purchasing decision.
Lisbon: I can understand where you are coming from, when you choose a platform you make a bet, though us here are mostly happy with the (work)horse we chose.
However if you need the security an open-source solution with community backing, perhaps the arduino world have the answer for you and of course the latest raspberry pi is just twenty bucks
Yes, I already have an Arduino. The imp seemed such an elegant solution - small, all the features I needed, great price.
I (incorrectly) assumed the imp cloud was an option for those that didn’t want to set up a local wi-fi solution. I must admit I’m a bit gutted now I know the imp cloud tie-in. I’ll use it for a bit in case any customers want some coding done and wait for someone to hack the imp to allow local access.
Like I say, however good the product, can you ever realistically assume that the manufacturer will always be around?
the beauty of the cloud is that ElectricImp can do a massive rollout of patches when a security issue is found. and if you write your code properly, local functions will continue to work when its offline for a few hours.
This is closely related to the fact that the Imp cannot make/receive local network connections.
‘Imp-like’ devices such as the SparkCore are open source and do allow you to run your own cloud and make/receive local network connections. While its nice device, its buggy. plenty of options to crash your core, requiring a powercycle before you can flash it again, if you are quick enough, because there are not bufferd firmware uploads. Also, im not sure if i want a device like that in a commercial product in my house, because you CAN make local network connections. i have no idea what the manufacturer is doing with it, now and in the future. or when they are hacked.
remcohn: Re: “the beauty of the cloud”
I get the usefulness of the imp cloud but none of the answers above have answered the question - what happens if the cloud closes down (eg the electric imp company ceases to exist)?
It would be useful for Hugo (or anyone at the company) to let us know what we would do in that circumstance. Just because the cloud is very useful (and I would use it and I would be grateful for fast patching) if it’s not there and “local functions will continue to work when its offline for a few hours” what happens after those few hours?
It would be good to know that in the case of electric imp going bust, they have something they could enable/release to allow a backup cloud/local server to take the strain till a permanent solution is found. Letting us know there’s a safety net if the worst happens (some details needed) doesn’t hit their revenue stream and it encourages take up.
This issue has been discussed in other threads. If electric imp were to go out of business, there is realistically nothing we can promise. Any promise we might make can be invalidated by a bankruptcy court; this is simply a fact of how bankruptcy works - any contract can be broken. It’s most likely that the court - or our investors, wanting to extract the most value from their investment - would want to sell the business as a going concern to another company, as it has income and an installed base, which would mean little or no service disruption, but to be clear that’s just logical speculation, and strange things happen in those situations.
Yes, some large customers have other options which would not be available to small customers, but that’s how business works I’m afraid: you pay more, you get more. If I had paid for a first class ticket and my flight was cancelled, I’m certain the airline would be doing a lot more to get me to my destination than someone with a discount economy ticket.
As I noted, and is pretty clear from everything we do, we are a service company. Commercial users pay an ongoing service fee. The service fee is not “a big unknown” - just contact email@example.com and they will send you the pricing. I know our hardware is great, but that doesn’t mean we are a hardware company; the real value we provide is the integration between hardware, OS and service, along with long-term management of service and security.
As paying for a service appears to be a sticking point for you, I would suggest you use a different product for your customers: I don’t believe this will be cost effective unless you value your time and skills at $0, but it’s absolutely your choice. There will never be an imp available which does not use our service, so don’t hold your breath for that one