This week at Mobile World Congress Barcelona, Facebook and BRCK are announcing the open sourcing of Magma, a software platform that contains the tools needed to deploy and extend LTE mobile networks in under-connected and unconnected areas of the world. This includes software powering the mobile packet core and network automation and management tools. The reason this is important is that it allows rapid deployment and software development on a technology assumed to be available only to the incumbent MNO’s, and only with software and networks provided by a few behemoth suppliers (Huawei, ZTE, Nokia, etc.).

BRCK’s 2,700 WiFi hotspots in Kenya and Rwanda confine Moja users to a radius of 50-100m, a fundamental limitation of WiFi networks. To give users wider coverage, BRCK has been piloting a low cost, solar-powered, data-only LTE deployment that extends the Moja network with blanket coverage across kilometers, with the potential to cover thousands of people per tower.

Our Moja platform is the answer for affordable internet where people can’t pay to get online.  The business model allows consumers to connect to the internet for free, leveraging their digital engagement to ensure that the bandwidth is paid for. LTE allows that signal to not just work when you are within range of a Moja WiFi hotspot in a restaurant, shop, or bus, but works to keep you connected wherever you are – at home, as you hustle, or on the go.

As we continue to expand the Moja network, WiFi continues to be the tip of the spear.  It’s a great technology with lots of bandwidth, an easy connection model, and a very low cost of entry. We are continuing to invest in our work providing WiFI in public transit and have recently invested in fixed access, having acquired the Surf network with over 1,200 existing sites..  However, LTE has one great advantage: a larger footprint of coverage. LTE operates on licensed frequencies (although there is work being done to change this) and this allows us to use high power LTE radios to transmit a continuous Moja signal over kilometers, not meters, direct to a user’s device, connecting many thousands of people who otherwise don’t have access to affordable internet.

BRCK has embraced Magma as the backbone of our LTE network. The technology is similar to a lot of the work done in the early days of BRCK, building high reliability networks in resource-poor settings. It distributes the compute load to the edge, allowing for better performance despite the lack of resources. It uses modern web-compatible APIs and communication protocols that allow us to integrate seamlessly into our existing platforms and, most importantly, it breaks the reliance on legacy systems that have put scaleable LTE out of the reach of most.  And it’s built on software defined networking, so our upgrade path to 5G isn’t limited by hardware or vendor lock in, just a small matter of programming.

We’ve done a lot of work internally to Magma and shared this back into the open source effort. We’ve built out an MEC (Mobile Edge Compute) platform, localising the Moja platform and increasing the speed to the consumer. We’ve built a network monitoring stack for our wireless backhaul links, and we’ve integrated the platform seamlessly into our WiFi offerings and mesh networks that we run. We’ve also been integrating radios from multiple vendors for different deployment scenarios, as well as making Magma work with our own SupaBRCK microserver.

BRCK’s LTE stack is built using the SupaBRCK at its core. SupaBRCK is a rugged, outdoor-ready microserver powered by an Intel x86 processor, with 8GB of RAM, and up to 5TB of storage. Using SupaBRCK and the Magma gateway on the tower lowers the capex and opex of the site significantly by providing a managed all-in-one hardware solution, and by moving all the critical control traffic off the backhaul network. Innovations like this allow for reliable connectivity even in last mile situations with a lower cost of backhaul than is achievable with other solutions.

Moving forward, we are actively piloting more sites with LTE. Deploying Moja on a low-cost solar LTE platform allows us to increase our coverage into underserved areas that traditional operators typically don’t see financial viability in. We are excited about solving the Last Mile Connectivity challenge for Africa and other emerging markets.

Increasingly, tech companies are moving to Shenzhen, Taipei, and Seoul. At the same time, BRCK is getting itself further and further into the remotest parts of Africa. We are a unique company, and we have a unique perspective on tech that makes life mostly hard, but ultimately fulfilling. We talk a lot about ‘You can do hard things.’ However, now we’ve come to the realization that ‘We make things hard’ is becoming just as true. But it’s our customers that do harder things.

Coming to Pemba has given the team here a glimpse of island life and the challenges that come along with it.

The goal today was to install a small weather station at the lighthouse on Pemba. We’ve installed these weather stations before in various places, including on Mt. Kenya, and this was a good excuse to test the new PicoBRCK V3. PicoBRCK is an IoT gateway and edge node. We’ve been working on this now for a few years, slowly perfecting what it means to connect sensors in some of the remotest parts of the globe.

The weather station we’ve used as the base is mostly unexciting, however, the PicoBRCK is. The V3 now includes remote OTA updates, onboard flash storage, and even lower power usage than before. But what we learned was completely unrelated.

Arriving at the lighthouse, Liz and Birir began to prepare the weather station. The team had shipped some new base firmware, so we needed to do a manual upgrade of the device before putting it online. First thing to do was to download all the tools to flash the firmware, so out came the SupaBRCK and within moments we’ve got the truck set up as a mobile hotspot, a quick skype call back to BRCK HQ and we are off again.

Lesson 1. Third party tools make life hard when you are alone.

We’re more privileged than most. Travelling with SupaBRCK has meant that we’ve had pretty constant access to communication, something that isn’t a reality for most here. Without the SupaBRCK, we would have never gotten the firmware updated, and the weather station would never be up. We need to make our offline tools work for those who are offline. We’ve been building online tools for a while, and in building those out, sometimes it’s easy to forget that you’re a long way from online a lot of the time in remote areas. Sometimes that little bit of information can help you get connected faster.

Once we had the system set up, it was about getting it up and reading real data. Mark, who’s spent a lot of time in Tanzania, was set to talk to Yusef, the lighthouse keeper, to get his blessing. Well, it turns out we didn’t have the approvals we thought we did, and we had to improvise to make it work.

Lesson 2. Flexibility is critical when you are expecting the unexpected.

When we got permission to mount the weather station, we had to work out how to mount it. One of the great things about PicoBRCK is its small size, but the weather station was bigger. We’ve been thinking a lot about different options for mounting the PicoBRCK and the SupaBRCK. However, sometimes in the office it’s too easy to think about standardizing everything, and trying to force our installs into simple boxes. Mounting the weather station involved everything we had in the toolbox, every zip tie we had and scavenging U-bolts from every antenna. Deploying with all eventualities in mind, and the ability for the hardware to be flexible is critical.

We loaded up the daladala’s again with all the gear and people and set off to the new digs, Verani Beach, our home for the next three nights. Everyone got sorted out, went for a swim, and chilled out for a bit. Brandon from the Good All Over team got out the drone and headed out over the ocean.

Splash!

All of a sudden Brandon bolted up and ran at full speed towards the sea, a few people looked up, but no one connected his urgency with the splash of moments ago. Minutes passed and finally the penny dropped that Brandon was chasing the drone as it bubbled towards the ocean floor. Six of us jumped up and chased Brandon down. ‘He’s got it… come the f#$k back!…’ repeated from the shore. We hauled the drone out, and Brandon, exhausted from 300m of race-pace swimming and countless repeated dives needed hauling out as well.

Lesson 3. Keep it simple stupid.

A DJI Inspire drone is an amazing piece of machinery, designed in Shenzhen, the greatest electronics and robotics community in the world. It wasn’t meant to be field stripped on the sandy coastline of Pemba, hundreds of miles from the nearest oscilloscope, but we knew we had only a few hours to strip it down and get all the salt water out of the complicated electronics, bearings, and mechanisms.

We carry a lot of tools with us, but we did not anticipate the 300+ screws of 10 different types, from M8 hex bolts, to M2 pozi’s, and M3 Torx drives. We were laughed at by the guys working on the Land Rover, with their 10mm spanners and a single screwdriver. Tonight we’ll attempt the recovery, without wrongly sized screwdrivers, and see if we can get this thing back in the sky for the trip home.

It’s been an interesting day, full of engineering, personal, and team challenges. We’ve succeeded in getting our objectives done, but we’ve learned a lot today about how to make BRCK products world beating in our part of the world. Why do we use technology designed in New York and London and Shenzhen for problems in Nairobi and New Delhi… and Shimoni?

Thanks goodness for BRCK. 

Unfortunately this will be my last expedition post this trip. I’ve had to leave the guys early to make a dash north to sort out some business. If it weren’t for BRCK I’d still be in Livingstone. When the news came that I had to come back to Nairobi, we knew we had 40 minutes to get flights. The BRCK was fired up, with in a minute we were online, and 4 of us rapidly trying to find the fasted route north. Its times like this that the utility of BRCK really shines!

Earlier in the day the BRCK became a lifesaver as well. On the road from Lusaka to Livingstone, a short 563km journey, the Land Rover developed a nasty fuel system problem with the fuel lines filling with air every 20km or so, stalling the truck mid drive. Stopping every 20km to bleed the fuel system makes for a LONG trip (14 hours I believe!). Out came the BRCK, and everyone who wasn’t driving was delving into Land Rover forums the world over trying to work out what the problem was, then trying twitter, facebook and instagram asking for help. Eventually diagnosing the problem as a faulty lift pump. Another quick Google search for a replacement lift pump in Livingstone and we were on the way again.

Trying to get help with a buggered 300TDi fuel system

Power issues again. 

We are an expedition company, and as such, our vehicles are kitted out more than most. We’ve got a 1500W 110V inverter wired in to the battery, 4 x 5A Cigarette jacks, and dual 12V batteries powering all the kit we have on the road. The biggest problem? Finding cables! Every time we have to charge a GoPro, power the Satellite internet or charge the motorbike headset, a mad scramble ensues trying to find the adapter, cable, or correct power BRCK. This is why we love the microUSB. As from Part I, we know its not perfect, but now we are torn. What should we use!

Knowing your limits

Getting the BRCK online in Tanzania and Zambia has been really easy, in fact, other than getting SIM’s and working out the crazy commands, the BRCK’s have been working flawlessly. One area that has popped up constantly is ‘how good is my signal now’. Driving from village to village the signal fluctuates quite a bit. We are working now to streamline the process of finding out how your BRCK is doing now.

Back in Nairobi we are working feverishly to take what the team and I have learnt to get some updates ready for the trip north on Saturday.

Working on new wireframes with information from the expedition south

Along this trip, we are really putting the BRCK to the test. We are trying as hard as we can to dog-food this thing in the most extreme environments to see where it falls down and where it stands
up. Here is just a snippet of what we’ve learn’t so far. Some of it reinforces what we know and a lot of it, thankfully, is breaking down our preconceptions.

“Pickin’ up good vibrations”

We did a lot to make BRCK robust, we’ve used thicker than normal circuit boards, everything is tied down with multiple methods and in all directions. Hitting these washboards, thowing the BRCKs in
the truck and just plain misusing it has shown a few things. So far its done well, with all our BRCKs behaving flawlessly, other than one small thing. The microUSB cable keeps walking out! We know
we need to move away from this connector, but we really love it for so many reasons. Its one less cable to lose, one less to carry, and I have 20 spares at home.

450km of this. #brckexpedition

A photo posted by Reg (@regorton) on Nov 11, 2014 at 9:30pm PST

Dust and Water (Fire and Ice?) 

Dust, its part of what defines Africa and part of what makes working in this continent so hard. The dust here is like nothing I’d seen before moving here, and something that is hard to comprehend.
If you look back in the blog, you may see our trip to Turkana. In this trip, we changed a lot about how V1 was made, specifically we moved the power button from the top to the side, and really
improved its weather sealing. Its these small changes that make a huge difference out here in the bundus. Driving from Arusha to Dodoma in Tanzania reinforced that thought. While the BRCK fared quite
well, it always astounds me where dust can get to.

We’ve often run on the theory that if you have a BRCK, you have a smartphone, laptop or some other high value equipment. Unfortunately this trip has shown that’s just not as valid as we hoped. The
BRCK is your often your only connection, your lifesaver, your redundancy. I have a laptop, a phone and iPad and many more, any one of these can get me online. I think the assumption is not true any more.

“Failure is not an option.”

But it will happen. We have to be realistic here. We are working in environments that are tough, dirty, messy and we must. As above, in order to make the device less complex we purposefully used
the thought pattern that ‘the BRCK will do the hard work for you’, but its frustrating to just wait in the dark. I want to know why it’s failing, and I want to help it get online now. We are activly working on improving the offline experience of the BRCK, like said before, its not going to be a traditional router experience, but conversational, something that gives you the information you need to know, but isn’t just for a networking expert to understand. With us on the trip is Mark Kamau, from the iHub UX lab, Mark is going to help us form a better mechanism for explaining why you damned BRCK isn’t working and how you can get it back online.

The Land Rover just 48 hrs before departure

“Right here, right now”

This trip has been taxing, and we are only half way though. Sometimes in these taxing times when you are tired, dehydrated and hungry the last thing I want to have do is fiddle with tech, all I
wnat to do is make sure things are working, send an email home and get to bed. Right now there are 2 speed hiccups with BRCK. Firstly boot time, the BRCK typically boots in less than 45 seconds,
but then latching a network takes time and if something goes wrong then trying to rectify that with antennas, APN settings, checking the charge level all take time. Also, just doing prechecks in
the morning takes time, there are many things to do. Check the oil, check the tyre pressure, GPS batteries, everything bolted down and . Right now, we are working on the offline functionality to
speed this up and give you more information right there. For V2 we are going to address this, additionally, we are investigating display options so that you can see the status of your BRCK even
when its off, getting you all the information you need instantly.