Custom/Turnkey SAFEnetwork farming hardware

I am currently pulling together a POC of an alternative to the “21” computer with SAFE farmer and an API for clients. I believe this will demonstrate a very compelling alternative with much lower power requirements whilst being able to offer very similar API’s (launcher is almost there by default) to the 21 computer. There is a model there and it could be satisfied very quickly. Looking at beaglebone black and potentially a partnership with a supplier, like adafruit etc. It’s all very much a side project at the moment as I am very pushed for time, but it is very doable and with not much work, I believe. I hope to post something within a couple of weeks of MVP.


Interesting that you’re looking at the Beaglebone, it’s got lower specs than I’d have imagined: 1GHZ ARM & 512MB RAM, USB 2.0. £55

I’ve been looking at the Cubietruck for a similar project 1GHZ dual core ARM 2GB RAM, USB 3 & Sata 2.0 £69.99

Do you think the extra £15 would be worth it for the 2GB ram & faster disk read/write? I realise the hardware isn’t fully Open Source, although much of it is It’s only the Wifi driver that needs a binary blob & that probably isn’t important to have working in this use case?


Yes it will do for sure and worth the extra £15. I think the BB is doing well with lots of capes etc. I am looking for a fully open source implementation though so am a little swayed by that, well for this purpose anyway. I think if we get the reference implementation open then specialists will of course do what you are doing here and beef that up a little. So in essence I think you are making a good choice for your use case here.


Phew! I guess the Beagle Board X15 1.5Ghz dual core, 2GB Ram, 3x USB 3.0 is one to consider as well although it looks like it’s going to be £165 which is on the pricey side


Hi David

Thats great! Something like a “21” computer is exactly what we need. The reason I mentioned Adafruit and probably why you are thinking them also is that they have a heavy focus on learning, tutorials and training which I think is a must. And I see they accept BTC, so are already one step ahead of other suppliers in my book.

I also think another requirement from the supplier should be the ability to add on and order extras.

Do you think we should post a poll to gauge how many of us would be interested in placing a pre-order if the price and tech is right? I think if we had some real numbers our discussions with a supplier might be a little bit easier.


I think this is a very good idea. I am interested in garnering thoughts and wishes here. I see a great opportunity to grow the network pretty fast and give developers something to really work with in the Iot space.


Are you also looking at this as a possible off-the-shelf farming option?

Or both options as something separate?


1 Like

It would be a great deal to have available a machine that is MaidSafe ready (capable) and if it’s an effective farming machine, I’d expect many would appreciate that being available. And certainly provide a benchmark for future hardware considerations.

And I recall how miners in the POW arena make purchase of many graphics cards at many hundreds in price points for that POW mining. I estimate there ought to be a marketplace for the sure settings of running safe_vaults. And especially at the onset of test_farming, and farming itself.

Perhaps even launch on kickstarter and introduce people to cryptographic integrity through the safe network while buying a machine expanding the horizons and outreach. With instructions that to obtain safecoin you can buy this machine and keep it running eventually there will be safecoin, and instruction on how to use it to store data.

Devs would have access to safecoin from the machine that they can be easily use as part of developing or likewise participate in the network; and even yet so, incorporate safecoin to their professional work, fitting safe network and its data storage capabilities with great security into the use cases of everyday software engineering work. some thoughts…


If anybody here creates a Kickstater campaign I will contribute funds in return for a unit. Or if a supplier can put something together I will pre-order :slight_smile:


Both I thnk …


Already looking that way for a big splash. Lets see what can be achieved in next few weeks. Waiting on a couple of Rust PR’s to land (multirust and build-bot) which will allow very easy cross-compilation. Theres a few things happening right now that should all come together very soon.


It would be really good if it could be done in such a way that it is easy to port to other cheap similar boards.

But a “off the shelf” unit would be great.

BTW: ever heard of who sell a lot of hardware to a wide range of people. Maybe when its a product that can be sold they might pick it up too.


Yes I am looking to create a reference platform and image. The idea being easily audit-able and transferable by anyone. This should help with security and also act as a reference platform for technology partners who may want to jump on board as well as community members creating their own personal or mini business as well. A few wrinkles to sort out but it is also a good test of the launcher API as well as @eblanshey API and more. Of course this is all late night when I get time, but seems easily doable and would provide some stimulus just when we need it. I think this can be initially partnered with the likes of adafruit sparkfun etc.(who seem to be nice makespace type companies with similar goals to us in many ways) My only reservation there is the USA link to be honest, not meaning to cause waves, but companies in the USA have their own worries at the moment. So we need to check how safe it would be and how to structure such relationships for all parties to be secure in themselves.


There is always who are in Asia and support opensource projects. I have bought from there (32 bit opensourced logic probes better than the commercial ones) and had no issues with their supply etc.


same here, count me in!


There’s certainly more people living not in the USA, so focusing on the rest of the world would help more people. And likely someone might come up with hardware supplies in the usa especially with a benchmark hardware already out there to use as a basis for the next thing Or partner up and import your creation to usa.


I was looking at the Radxa Rock series (OSHW) as an option for a Freedombox. Perhaps it would be worth considering, too?

1 Like

Probably hijacking your thread @dirvine, but I wondered whether it is feasible to port the linux version of the future SAFE client to Synology´s DSM. I am not a technical guy but from what I see that shouldn´t be all too complicated and it would make it easy as pie for custom users to get started with the network. With the DiskStations thousands of users already have potent hardware to support the network.

Of course if it´s a precompiled version i.e. by Maidsafe, most users here wouldn´t consider using it. However, if we think about it as a strategy to market SAFE to end users with capabilities to provide ressources to the network, I could imagine this as a very productive approach. Without knowing it, I guess it should be possible to compile the client on DSM as well (at least I don´t see why it shouldn´t work). What do you think about that?


It may be, I feel the reference build should be a roadmap to follow for many systems. There is a couple of things in rust right now to make this super simple to port. So there is a chance we can make this very much a reference install and guide as well as a product in it’s own right. When we can target headless arm devices then I think we will have a wide audience available to us. Systems like DSM would then I imagine find it even easier to integrate. We will see though. I think this may become a project in it’s own right with a team associated with it if we want to do it right. A lot of work but looks like low hanging fruit to me.


100% agree! I have loads of TB ready to throw into the network :smiley: