Towards a Democratically-Distributed IoT

The questions raised in my previous post are worth serious consideration.  If the commanding heights of global industry are to be conquered by robots, then this is a step towards a distopian cyberfeudalism straight from the pages of science fiction.  To the uninitiated, the vision of robot overlords may represent something with some resemblance to a human, like the Terminator, or even the 1980’s toy the Omnibot, complete with the computerized voice and choppy, robot movements.

In reality, these robots are much more likely to look very similar to things we already see in our everyday environment: cars, tractors, pumps, electrical outlets, automatic fences, etc.  The essential difference will be one of connectivity and control systems; cars and tractors, instead of acting as discrete units on the landscape, will be interconnected, receiving data from and sending data back to the server farms of the Internet cloud.  Hardware interconnectivity will be mediated and controlled by software, which is the area of greatest development at the moment.

Development and deployment are the clear and present opportunity for the open source entrepreneur.  Google, Facebook, Amazon et al. (ie “the Stacks”) want sole and exclusive access and control of the hardware, software, and data of the IoT.  And while they certainly have the capital and the clout to make that happen, the final outcome is hardly a fait accompli.

Nevertheless, I am left with numerous questions that, if answered, may provide some insight into the path forward for building a decentralized, P2P IoT that is owned and operated in a more democratically distributed fashion:

  1. What are the elements of successful open source projects?
  2. How can these projects be funded in a business environment largely driven by venture capital and promises of multibillion dollar returns?
  3. How are these projects coordinated over large, disparate groups that all have a stake in its success, but may not be able to lead their development?
  4. What is the role of the worker-owned cooperative, or coalitions of such cooperatives, acting as decentralized groups of consultants developing and deploying systems in the real world?
  5. How do we bridge the gap between technologists and industry specialists (farmers, construction workers, doctors), such that developed technologies reflect and address the needs of realities on the ground?
  6. What real and immediate monetary incentives exist for technologists to develop and deploy these systems in conjunction with industry professionals?

All of these questions point to a development pathway that is bumpy and unpredictable, but also full of opportunity and potential.  If you have answers to any of these questions, or feel that you would like to see some more added to the list, please feel free to chime in on the comments thread.

IoT and the Future of Open Source

Photo courtesy Lima Pix

Photo courtesy Lima Pix

The Internet of Things is here, in a fledgling state, but arriving with all the speed and disruption that the worldwide web and Web 2.0 did in previous decades.  With the Internet of Things, everyday objects are now connected to the global network, providing data on everything from ambient temperature in a domestic bedroom to the temperature of a nuclear reactor core in a utility-scale electrical power plant.  The McKinsey Global Institute estimates a market size of $6.2 trillion by the year 2025.  According to an article by the the Siemens corporation, “The market research and consulting firm International Data Corporation (IDC) estimates that 32 billion objects will be connected to the Internet by 2020, and that these will produce ten percent of all the data generated worldwide”.

Of course this increased influx of data will be broken down along traditional industry lines: there will be an IoT sector for residential construction, for agriculture, healthcare, and manufacturing.  These systems will likely be a blend of proprietary and open source solutions, stacks of small embedded hardware, coupled with sensors and actuators, streaming data to the cloud for processing, storage, and analysis, then acting on that data through various control systems like heating and cooling, water pumps, and medical devices.  In some cases these systems will greatly enhance efficiency and productivity, allowing for further mechanization and the greater integration of electromechanical devices (ie robotics) into industrial processes.  In other cases these systems may add greater capital expenditure via deployment and maintenance costs accompanied by insignificant ROI; these systems will fail.  These are all things that will play out in the IoT marketplace over the coming decade. There will be winners and losers.

At first blush, open source software and hardware seems poised to be a big winner in the realm of IoT.  Ubuntu is currently working on Snappy Core, “an app store for open robots.”  And just recently Google announced Brillo, what they are calling the “underlying operating system for the internet of things”, which will be Android-based and at least have some open source components.  Arduino microcontrollers and similar open embedded systems are going strong and continuing to grow in popularity, versatility, and diversity.  Is open source software and hardware on the verge of taking over the world?

Not so fast.  For the mega-corporations bringing these products to market, open source may be a convenient development approach, and may have a familiar appeal after years of greenwashing faux environmental initiatives, but it bears no resemblance at all to Richard Stallman’s philosophical stance that “non-free software is a social problem and free software is the solution.”

While open source flourishes on the cloud, in your Android handset, and even in unexpected places like Facebook and Microsoft, the smartest of the proprietary mega-companies have long since realized that the true value lies in the data sets that end-users generate while interacting with the software. An example of such interaction with the software can be observed on the site, which stores the data of the best online casinos. A large number of users on the Internet choose Betsson-casino for interesting content and unique bonuses of popular casinos. There is also a mobile version that is available from any Android device.  Facebook and Google are poster children for the new corporatizating of private data; but now as these data tracking systems move into the physical world, tractors are spying on farmers, sending back their data so that georeferenced industry-specific knowledge can be aggregated and analyzed, with the end-game being the complete robotization of the factory farm.

The question remains:  who will own the software and hardware systems as they are deployed into the world in the coming decades?  Clearly, however, it doesn’t just matter who owns the hardware-software, but ultimately and perhaps most importantly, who will own the reams and reams of data that will begin to flow as the Internet colonizes the planet?


Photo Tweets: The Crops of Peru and Bolivia

Qañiwa is a close relative of quinoa, but lacks the saponins that complicate the quinoa post-harvest processing.

Tweets from Bolivia

Here are some Tweets and associated images that I have put out over the past few weeks from Bolivia. Unfortunately Internet connection hasn’t been consistent enough to put out much content, but hopefully will be able to do a little writing in the days and weeks ahead.

Owly Image

Tune In to the Food Tank Summit Live Stream

Food Tank SummitThis week’s episode of the Agroinnovations Podcast features Danielle Nierenberg of Food Tank to talk about the upcoming 1st Annual Food Tank Summit.  The Summit will take place in Washington DC on January 21st and 22nd.  Although the summit is sold out, you can participate via twitter using #foodtank, and you can also access the summit online via an ongoing live stream.  Please take note and participate in this great event.  If you do participate in this event, please share your experiences with the Agroinnovations community.

Software Defined Radio using GQRX on Ubuntu Linux

I’ve been playing around with Ham Radio quite a bit lately.  The potential for varying applications in agriculture is pretty broad, and it’s just plain fun.  Here’s a little demo I put together of a very cheap Software Defined Radio, which will be a big growth area in coming years.

Just Wait for the Collapse

Fremantle Bridge Collapase July 1926.

Fremantle Bridge Collapase July 1926.  Courtesy Wikipedia, photo by Donna Barber.

Seems like there’s a new affliction amongst some segment of permaculturalists these days, maybe we could term it collapsicosis or collapsitis.  This peculiar and apparently growing meme goes something like this:  We in the permaculture community are doing everything right, and the evil people who care about evil things like money and competition are destroying the world with their globalized corporations, fiat money, agricultural subsidies, and cheap oil.  Trying to compete with or displace their system is a wasted effort, and certainly using any of their methods for converting resources into products and services is akin making a deal with the devil and will taint the moral purity of permaculture.  Anyway, this whole system is on shaky ground, billions of people are going to die, so let’s just wait for the fiat money house of cards to crash and burn, and then we can get around to the real business of doing permaculture.

What I am proposing, however, are permaculture enterprises that are scalable based on pooled efforts and resources.  As I have already said, economies of scale are basically natural law.  Failure to execute is our own fault.  Lack of good business and economic data, our own fault.  A scarcity of polyculture input-ouput models, our own fault.  Undercapitalized permaculture enterprises…yeah, that’s partly our fault too.

I am offering a concrete solution that reflects the reality of the world we live in today.  And this solution is directly within your locus of control right now.  We don’t have to ask permission, get a government grant, or wait for the collapse of fiat money to create worker-owned permaculture production cooperatives.  Any other arguments that are not based on strategies of direct social action are only so much white noise.

Understanding Economies of Scale

This week’s podcast epsiode, the Failure of Permaculture, has created a lot of debate and some controversy.

A first point is expressed in the below quote from @ajtarnas.

“The podcast doesn’t include any case studies of permaculture or agroforestry projects failing, it shows cases of small farmers (two of whom were hobby farmers) failing. Bren Smith would definitely qualify as a perennial polyculturalist, but he doesn’t mention it at all”

True but the point is still illustrated by the case studies I provided: homestead to small commercial scale agricultural production is a huge financial challenge.  Can it be done profitably?  Sure, people do it, but permaculture is chronically capital starved.  Why?  Bren might not mention permaculture, but he is practicing it.  And I’m sure he works with others who practice it too.  And they are all struggling.  We could find plenty of cases of permaculture farms who face the same constraints as the hobby farmers.

And again I come back to my central (but evolving) thesis: permaculture is failing because we are only practicing one part of it effectively.  Yes, we have got down pretty good the polycultures, the soil biology, the high density grazing, and the seed production.  Not across the board, and there’s a lot of scaling up to be done, but we have good replicable models for working with agroecological production systems.

But our understanding of economics is poor.  Let us take the time to define an economy of scale.  According to Wikipedia:

economies of scale are the cost advantages that enterprises obtain due to size, output, or scale of operation, with cost per unit of output generally decreasing with increasing scale as fixed costs are spread out over more units of output.

Economic models of economies of scale are supported by many years of data, the laws of the physical world, and common sense.  The graph below illustrates the point, as enterprise size increases, average cost per unit decreases coupled with a commensurate increase in productivity and volume.

Image courtesy of Wikipedia.

As quantity of production increases from Q to Q2, the average cost of each unit decreases from C to C1.  Image courtesy Wikpedia

As economies of scale are achieved, the enterprise benefits from many of the other advantages of size:

  • An increasingly specialized workforce, which in the case of permaculture is critical considering the broad range of expertise required
  • Improved negotiating power with buyers and sellers
  • Access to large volume purchase and sales contracts with other complementary enterprises, and therefore more favorable product prices

And then this comment from the thread on reddit:

“Of course, the ‘small permaculture farm’ model is economically less efficient than big agro-industrial farms, less adapated to today’s financialized economy.”

My counterpoint is that we’ve done this to ourselves.  Nobody’s fault but ours.  Who said permaculture farms have to all be small?  We have not fully understood that Bill Mollison had a view of permaculture that included finance and economics, and we have mostly ignored that to our peril.  If we that were not the case, we would see hundreds, if not thousands of large permaculture enterprises creating employment and regenerating degraded landscapes by understanding and leveraging optimized economies of scale.

And here another quote, again from the reddit thread:

Don’t you think that the issue is not permaculture having an economy of scale, but that the current market created by subsidy and a permissive attitude towards externalities creates an environment in which only very specific approaches to farming are economically viable?

Yes, I do think that the regulatory environment makes permaculture difficult indeed, but if we were creative permaculture designers we would design our way out of this problem.  Like I said, economies of scale are basically a law of nature; as a rule, we are not incorporating this in our permaculture designs.  We are supposed to be observers of our world, and model our permaculture accordingly.  If we want to produce more, access more land, become a bigger economic and political player, then we have to implement some form of economies of scale. Even if you personally, reader, are not interested in participating in such an endeavor, I would hope that you still support the concept as an essential part of a permaculture grand strategy.

The Permaculture Fail

Update: This comment thread on reddit for the podcast episode is red hot right now.

This is an excerpt from this week’s podcast where I argue that permaculture is failing.

Sustainable agriculture faces a crisis.  In many ways this crisis reflects the broader social and economic breakdown crisis that American families are facing today: economic hardship, social inequality, environmental degradation…all of these are reflected in 21st century agriculture.  So now reader, I ask you these questions: Why did you get interested in organics, permaculture, and sustainable agriculture in the first place?  Was it because you felt you could be a part of transitioning agriculture into a new and sustainable model?  Or because you romanticized about the agrarian traditions and lifestyles of a bygone era?  Was it to prove a point, that there is a better way?

Or perhaps you were born into a farming family, and wanted to carry on the legacy, like Kasha Bialas.   Bialas describes the life of the 21st century American small-holder:

I’m a single mom with a fifth grader to wrangle and I spend the bulk of my early morning and evening hours at the computer organizing our CSA farm share program, developing newsletters, making website changes, creating advertising fliers and recipe handouts, and occasionally doing the mom thing. My days are spent doing all manner of farm work.

I will also ask you this: Do you think we are winning?  Does the portrait Kasha paints, of lone crusader, struggling to keep her head above water, sound to you like a movement triumphant?  Let me draw your attention once again to last week’s episode of Agroinnovations with Dr. Joe Kovach.  While Dr. Kovach has demonstrated the possibility of earning $90,000/acre on a small scale, this demonstration comes with a number of critical caveats that Kovach himself identifies in the course of the interview.  I have summarized them here for you:


  • The price discovery mechanism in the farmer’s markets conveys a marketplace of small producers competing with, and undercutting, one another instead of competing with supermarkets and industrial agriculture.
  • Kovach also states: “I don’t see how anyone makes any money in a farmer’s market.”  He attributes this to very low sales volume relative to production and retail labor.
  • Some of the techniques he describes, like season extension, increase profitability for the individual farmer through offering normally non-available produce during certain times of the years at higher prices.  Yet this competitive advantage evaporates when others adopt similar practices.  In other words, individual tactical advantages do not necessarily translate to a viable alternative business models for the sustainable farming movement.  Again, this is intra-movement competition.
  • Finally, Kovach says: “It’s a lot of work [to set up a polyculture like this].  The work is the deterrent.  The other major deterrent is you cannot compete with the major grocery stores.  Because our economies of scale and our agricultural system is so efficient, you can’t use polyculture to feed commodity buyers.  You can’t do this on 5 or 10 acres. It’s tough to do it on an acre.  But it’s also tough to compete with cheap oil.  You can’t compete on price until gas skyrockets, which it may sometime, but now that is what makes it tough to market.”


To be sure, I have great respect for Dr. Kovach and many others who are helping to show us the way.  Nevertheless, these constraints, and the many others that I have shared with you today, paint a bleak picture indeed for the future of sustainable food.  I ask you to consider the possibility that the permaculture movement is failing.  The reason for this failure is that we have focused all of our energy on biological production techniques, many and most of which are sound, effective, and replicable, yet we have done so on top of a broken socio-economic model.  If permaculture, or holistic management, or biodynamics, or any other such production or even decision-focused technique, was so effective, why then do we hear story after story of young farmer’s struggling, going into debt, working ungodly hours day after day, year after year?  The only people with any peace of mind are those who have made some enormous sum of money in other endeveavours and have adopted farming as a lifestyle or a hobby, or the rock-star Salatin’s of the world who make a good portion of their money on speaking tours and books.

I write this not to be discouraging or defeatist, but to impress upon you that it is time we started creating the socio-economic models that will make permaculture successful.  We have many options at our disposal, including worker-owned cooperatives, labor unions, and collective bargaining agreements.  It is probably fair to say some of these models have yet to be created.  I do believe that permaculture could one day provide a substantial portion of our food supply, but only if and when we begin to work together collectively.  The model of the rugged individual crusader, working her farm into the late hours of the evening, needs to be abandoned, as it has proven to be unmanageable and ironically enough, unsustainable.  In its place, we must forge a new model of collective democratic action.

Trends in Ag: Consolidated Capital vs. Open Source Innovation


Image courtesy

Recently I was interviewed for a CNBC article by Mark Koba entitled “Frankenstates: Winning the Agriculture Tech War.”   In it, Mr. Koba has the following passage taken from our interview:

While most tout the progress that technical innovation has brought and will bring to agriculture, there are still words of caution about its impact.

“The trend for things in agriculture is to get bigger and more consolidated, and that’s creating two markets,” said Frank Aragona, CEO of Agricultural Innovations, an information source for agricultural strategies.

Aragona, who has a masters degree in forestry, said the worry is that large farm operations can more easily afford the new high-tech advances, while smaller farms can’t, creating a technology gap of sorts.

“A farmer may have a tractor that’s 20 years old and now outdated,” he said. “But many farmers are saying it’s too capital intensive to go out and buy the newer, more advanced models.”

While the quote illustrates a recurring theme in our interview (the increasingly capital intensive nature of industrial, commodity agriculture), it does not provide the full context of my comments.

Commodity agriculture is growing ever bigger, pushing the boundaries of yield per acre, increasing the need for industrial scale inputs and machinery.  This has created the “get big or get out” dynamic experienced by so many farmers, particularly in the United States.  Commodity agriculture requires bigger machines, bigger science, bigger markets, bigger acreage, and ultimately, more consolidation.

What was left out of the article however, is the countervailing trend of small.  This is a trend largely driven by consumer demand to purchase food that is sourced locally, produced sustainably, and built upon a relationship with the farmer and her team.  To be sure, the percentage of the nation’s food coming from this economic model is still minuscule, but rising, and with it rises a recognition of the importance in supporting alternative models to industrial agriculture.

This trend extends into the realm of technological development.  Rather than waiting for big ag and corporate America to invest in appropriate scale technologies for the small and medium-sized farm operation, we have instead opted to build these technologies ourselves.  An open source ethos accompanies this can-do attitude, as we seek to build the hardware and software that will assist producers with a range of farm activities and challenges:

Farmhack is an organization acting as an umbrella for many of these activities, but other projects have also made great progress in the development of open source agricultural tools, notably the Open Source Ecology Project.

Together we are developing a set of tools that are freely available to all, with open hardware specifications built on top of open source software code with freely available documentation.  We believe that appropriate agricultural technology should be a part of the commons, much like the natural resources on which sustainable food production depends.  Therefore we are building a collaborative network to develop and deploy these technologies around the country and around the world, with the hope that freely available, open source, appropriate technologies will help us confront the many social, economic, and environmental challenges of 21st century civilization.

How to Attract Beneficial Insects


Attracting beneficial insects is simple. Plant flowers! Lot’s of them, everywhere, of all different kinds. Here is a lady beetle on a daisy flower in our garden. And if you see aphids, white flies, or other pests on your flowers, don’t kill them! Soon enough the lacewings, ladybugs, and others will come along to clean up the mess.

Food Price Inflation and Demand Elasticity in the Developing World

A twitter conversation with @ajtarnas has sparked me to write this post.  The conversation started with this tweet, quoting and linking to an article by AFP:

@ajtarnas responded:

The conversation concluded with this exchange:

So, in response to the comment that there is no context provided on this issue, and that food price inflation only affects people making less than $2 a day, I offer the following quick analysis. First a brief primer on the concept of demand elasticity. Elastic demand is when the percentage change in the quantity demanded exceeds the percentage change in price. Inelastic demand is the opposite, that is when the percentage change in the quantity demanded is less than the percentage change in price. The case of food prices will help to clarify this concept.

In the United States, as @ajtarnas points out, demand for food is relatively inelastic. That is, a 4% change in prices will decrease demand by less than 4%. Consider, however, that in most developed countries less than 20% of household income is spent on food. As the percentage of household income spent on food increases, demand tends to become more elastic. In India, 56% of household income is spent on food, in Tanzania 62% of household income is spent on food. In these cases, a 4% increase in food prices may result in 5% or greater decrease in demand; not because people don’t need or want food, but simply because they can’t afford it.

This phenomenon does not only play out from country to country, but also from household to household. Under current market conditions, the primary factor that drives access to food is income. What seems like small price flucuations to households high on the global income ladder are huge price fluctations to those on the low rungs of the global income ladder. One can easily imagine a family in Tanzania with more than enough money to eat a healthy diet. The social safety net in the United States, in the form of food stamps and subsidized housing, helps to mitigate the more extreme examples of malnutrition and starvation that are commonly witnenssed in Africa and India.

Even in the United States, poor nutrition is typically associated with low income families. So while these families may not be starving in the same way as those in Somali refugee camps, they are more vulnerable to obesity and poor nutrition as they seek low-cost substitutes to healthy foods.

Volatile food markets have played a big role in social unrest over the past 7 years. The impacts of climate change, energy scarcity, unstable financial markets, and political corruption will continue to exacerbate an already precarious global food security environment. While global commodities like corn, rice, wheat, and soy, will likely remain fairly abudant and relatively cheap for many years to come, other key ingredient to a healthy diet, like meat, fruit, milk, and vegetables, may become more expensive, especially for those functioning in the difficult world of elastic food demand. Fortunately, these products are relatively simple to produce for oneself, thus the permaculture solution is a viable way to supplement global human nutrition.

The Best Time to Plant a (Fruit) Tree

Red Gold Nectarine.

Red Gold Nectarine.

The best time to plant a tree was 20 years ago. The second best time is now. — Chinese Proverb

We planted a Red Gold Nectarine and a Summer Crisp Pear. Varieties that are both well-suited to our climate, and they were grown by our local providers at Tooley’s Trees.

When choosing and planting a fruit tree, be aware of the following factors:

  • The number of chill hours the variety requires
  • Be cognizant of the landscape location where the tree is planted; stay away from South-facing slopes, as your tree is likely to break dormancy too soon.
  • In our case the pear flowers later than the nectarine, so we put it in a sunnier winter-spring location. This way breaking dormancy early is less-likely to result in fruit loss from spring frost.
  • Don’t put a bunch of soil amendments in the hole, as the tree will not spread it’s roots into the poorer surrounding soil
  • Keep the first branching lateral roots close to the surface.
  • Spread the roots out in all directions in a mat-like structure surrounding the tree. Imagine a tree canopy that expands in all directions; you want the tree to replicate the same pattern underground.
  • Lift the upper roots as you bury the lower roots, then release and bury on top of lower roots.

Measuring Electrical Loads

Owly Images

Just got myself a P4460 Kill-A-Watt EZ Energy Meter that measures electrical loads from a standard 120 VAC wall outlet. This is a nice device that can help home owners find those pesky inefficient loads, and is a great energy conservation tool. It also allows you to put in a Kwh rate ($0.09/Kwh where we live) and then will calculate your running cost for a specific load on the fly.

This device is also useful for solar system designers. First, it can be used to help homeowners identify inefficient loads so they can take some conservation measures. This is a good step before sizing a solar array, as it can often reduce the total number of panels required. Generally, inefficiencies stem from household appliances like refrigerators and space heaters. Phantom loads also create small but insidious energy “leaks” that can add up over time; common phantom loads are cell phone and laptop chargers.

After taking conservation measures, the designer can use this tool to do a complete load analysis for the household or business. Generally, this approach is uncommon for grid-tied systems, as the local utility will provide average daily consumption information that designers can then plugin into their equations. For off-grid systems, however, this tool can provide an invaluable source of information to help gauge potential load requirements. A small tool like this is probably most useful for the typical homeowner or for the small business, but is less useful in large commercial and heavy industrial environments. And at $23 on Ebay, this isn’t a bad investment at all for the energy-concious homeowner.

Farm Hack: Fixed-Wing Drones for Farm Monitoring

Courtesy and Dorn Dox

Courtesy and Dorn Dox

Over the past several weeks I have been corresponding with Dorn Cox of about the uses of fixed-wing unmanned aerial vehicles (UAV), more commonly known as drones, for farm monitoring applications. Most people who follow this topic immediately think of quadcopters when they think of small remotely controlled UAV’s, but fixed-wing models have a number of features that are advantageous.

First, the most immediate benefit is the flight time when compared with quad or hex copter models. A quad is likely to get between 15 and 20 minutes of flight time on a single charge. Naturally, running all those motors at high speeds drains battery charge quite quickly. A fixed wing craft like the 3DR Aero has a flight time of 40 minutes on a single charge; one can cover signiciant territory in 40 minutes.

Currently fixed-wing craft can fly a path of GPS waypoints with much better precision and efficiency than a quad; this may change as flight controller hardware and software improve, which is happening quite quickly. Below is a sample image, which is quite comparable to the aerial imagery most remote sensing professionals get from more traditional aerial photography.

Aerial imagery is comparable to more traditional aerial photographs.

Aerial imagery is comparable to more traditional aerial photographs.

3D Robotics is now offering the 3DR Aero for a fairly reasonable price, though certainly you would have to put this rig to work to get a good return on investment. You might be thinking what I was thinking when I first saw this rig: a styrofoam frame! That will get wrecked in five minutes. But Dorn assures me that this stuff is pretty sturdy: “The EPO foam is tough stuff, can be repaired and is like flying your electronics in a shipping package- the sky walker airframe they use is less than $200 and some of them are less than $100. Cost is all in components”.

Before you go out and buy yourself one of these rigs, don’t fool yourself into thinking that this process is easy. According to Dorn, there are three areas that will require a signficant time investment and learning curve: the navigation software, the image processing software, and most importantly, learning how to fly. Fortunately for fixed-wing craft, there are some pretty good flight simulator software packages out there. Here’s a pretty good forum post that compares RC flight simulator software.

And remember, this technology is still in development, rapid development to be sure, but development nonetheless. There will be factors that you may not have thought of, like the nose shifting up slightly during stable flight, which affects image quality, or the hours required to get navigation software and camera to communicate, or the fact that you have to learn how to use CHDK (Cannon Hack Development Kit) if you want to have any chance of making this work. And of course, getting the images is only the beginning, then you have to process them and get good quality data. Samplepoint software is a good option for getting some usable data in consolidated fashion, though most aerial Samplepoint applications I’ve seen are still fairly experimental. The good news, again, is that development is happening at breakneck speeds. If you’re adventurous and technically-minded, you might want to dive right in. If not, you may want to check back in six months to a year and see if the technology has matured enough to make it completely plug and play.

Verizon Wants to Build Agduino, Too

Verizon spent millions of dollars on this ad campaign, considering it ran numerous times during the NFL Playoffs and the Superbowl. Agduino is a huge growth industry, but who will control the destiny of information intensive agriculture? Corporate America, or the OSAT hordes?

The race is on.

Drought in America: US Cattle Herd at Lowest since 1951


Texas Longhorns. Icons of the American beef industry.

I’ve seen numerous headlines over the past few weeks warning that the U.S. cattleherd is at historic lows.  Most blame drought for the decline, but looking behind the numbers a bit I realized that drought isn’t the full story, though certainly a huge part of it.

According to the University of Missouri, 2014 estimates for the U.S. cattle herd come in at 87.7 million head, down from 90.7 million in 2012, and continuing a downward trend that began in 2007.  Consider the following points:

  • In 1951 U.S. population was 154 million, less than half of today’s 317 million.
  • The decline in the cattle herd began in 2007, well before Texas was hard hit by drought.  The driver here was increased feed and energy costs, which greatly increased production costs for all links in the value chain, particularly for feedlots.
  • The process accelerated when Texas and other important cattle producing states were hit by a severe and prolonged drought starting in 2011.  Producers responded by feeding hay (which skyrocketed in price), moving herds to other non-drought states, and destocking.
  • The average beef cow herd in the United States is 40 head, which means many of these folks are diversified, small producers who may not have the capital or the energy (considering most are over 50) to get back into the game.
  • The cattle cycle runs 8 to 12 years, meaning even under ideal conditions it would take considerable time and effort to restock.
  • Conditions, however, are far from ideal, as drought persists in Texas and now California, an important beef producer in its own right, is experiencing record drought.

The bottom line is that beef is going to get more expensive, by as much as 4% in 2014.  In economic parlance, this means that supply should increase as producers respond to higher market prices.  However, as we have seen, agricultural economies can not respond like other industries (e.g. manufacturing) and ramp-up production at the flip of a switch.  There are too many variables at play, many indicating that beef prices may stay high, and the downard trajectory of the cattle herd may continue for many years to come, particularly if drought persists and as the aging demographics of the ranching community play out.

This portends big potential changes to the U.S. food supply in the years and decades to come.  A likely shift I foresee is younger farmers on smaller landscapes producing alternative non-ruminant protein sources, including poultry, fish, mushrooms, and even, perhaps insects.

Agduino Runner-up: Apitronics



The title here is not to suggest that Apitronics is inferior to the ManyLabs/SODAQ stack. Apitronics is built around the Beagle Bone Black, which was featured in a previous post about Agduino. Apitronics doesn’t use Postgresql, instead the system uses CouchDB, which may interface cleanly in data transfers to PostGIS, but I’m not sure. In general, I get the sense that Manylabs may be a bit further along in the process, but this is difficult to gauge with the information I have available. It would be nice if some of the disparate groups could begin to find a way to merge projects under a common umbrella, at least for the software development piece.

An interesting component of Apitronics is the “Hive” approach.
From the Apitronics website:

Our architecture is extremely modular consisting of Bees that can be “plugged” into any sensor and actuator, allowing the same device to be a weather station, an irrigation controller, or a control system for a greenhouse. Multiple Bees are coordinate [sic] by a single Hive which is a gateway device plugged into the user’s router. This device is a local coordinator for the entire network and preserves functionality without internet connectivity, making the Apitronics platform more robust and reliable than other cloud-based solutions.

I’ll be curious to see how this approach plays out in practice. Overall, Apitronics seems to be a solid project with committed project leaders. You can follow the primary lead for this project on twitter @thierylouis.

Agduino: The Top Candidate

The Solar Powered Data Acqusition Board (SODAQ)

The Solar Powered Data Acqusition Board (SODAQ)

So, after much searching and discussion, I think I’ve discovered some of the top candidates on which agduino could be built. I will share these in a series of posts over the next few days or so, starting with the top candidate. Thus far, the number one stand out is the consortium formed by Manylabs, Seeedstudio, and the creators of the SODAQ board. This project certainly deserves the support of agduino enthusiasts.

Before we get into the details of their project, let’s outline some of the elements needed for the successful design and deployment of agduino (in no particular order):

  • A comittment to open source at both the hardware and software level
  • The ability to connect with existing open source GIS software, likely PostGIS, which means Postgresql as the database engine
  • A physically robust and well designed circuit board, to deal with issues of reliability, durability, modularization, ease of use, and processing capability
  • An already somewhat mature software platform that is released to a broader community of supporters

Manylabs is focused on creating sensor networks for educational and environmental monitoring purposes. They are the driving force behind a large software development effort which will bring us a stable database/sensor/arduino interface. The end-user will be able to configure custom sensor networks, build relay controls, and collect and process data with no need to write code. Manylabs is building the system using a Postgresql database backend, with every intention of tying the system into PostGIS in the future. All the software and hardware is open source.

Seeedstudio is the primary hardware provider, and they have a pretty big selection of lot’s of arduino add ons. Have a look at their website, lot’s to see.

Finally, the SODAQ board is a solar powered data acquisition board which is still in development. It doesn’t use SD cards to store data (which are notoriously unreliable), has built-in memory, and can be powered by low-powered solar panels in remote locations.

Stay tuned as I continue to break down some of the other great agduino candidates.

The Agduino Roundup


The Beaglebone Black

Through lot’s of different conversations with people, I have started to put together a set of links that tracks some of the hardware/software stacks that are likely great candidates to work with agduino.

Most of us working in the area believe that a great deal of the software code already exists, but likely has not been integrated into a complete package. At it’s current iteration, this list is primarily focused on hardware. See comments below to understand some of the context around each hardware unit. If you feel something should be added to this list, leave your comments in the comment thread for this post, or if you prefer contact me.

Catkit: For connecting arduino sensors via ethernet

Comments: Cat5 vs. wireless is a pretty big debate right now. The Cat5 advocates say that power and data link can be provided over Cat5, and the implementation is simpler and involves less software and/or configuration. Wireless proponents believe that wireless is cheaper and easier to deploy, with greater flexibility.

My own opinion is that we will likely use a combination of the two, with great variation across production models and landscape scale. In hydroponics, greenhouse, aquaponics-type systems, there seems little reason to use wireless. On larger scale landscapes (e.g. a Western cattle ranch), we need to use wireless technology. Understanding radio frequency characteristics will help us to understand the best transmission frequency, which is likely not 2.4 Ghz (the standard transmission frequency of Wifi). 900 Mhz has a much greater range, with potentially lower power requirements.

Ninja Blocks


More easily write interfaces to customize the control systems for the arduino.

Comments: I can’t comment on Ninja blocks directly because I’ve never used them. However, an immature software architecture is the biggest bottleneck to agduino right now. So a coding system that allows non-coders to build interfaces could be huge, because it will greatly accelerate adoption rates and actual real-world application.

Reef Angel

A complete system for managing indoor saltwater aquariums. This is probably the most robust open source agduino-like community that I have come across. It’s built around a commercial platform, but the business model seems primarily focused on hardware sales. They’ve figured the manufacturing piece out, and they hired a good coder to write solid open source, cross-platform code. Reef Angel could likely be used for other applications, particularly aquaponics.

Arduino Tre


Comments: One of the nice things about the TRE is that it’s manufactured in the United States. In the great State of Texas. The TRE will expand the available code libraries for the microcontroller, because it will be able to use different coding languages besides just C+ (i.e. Java, Python, PHP, Perl). It makes a lot of sense to find ways for the microcontroller to run its own code, and to tap into the huge software libraries that are available from some very mature FOSS projects. This development could unfold very rapidly.

See my previous post on the topic

Beaglebone Black

Comments: Similar to the TRE above; this is ready to load Linux OS and go. Runs Ubuntu, Android, Angstrom (a Linux build exclusively for embedded). Nice. More on this to come.

Unite Arduino and PI

Comments: Fairly self-explanatory. Another attempt to turn the duino into a mini-computer, like the Tre and the Beaglebone.

Grove Pi

Comments: A modular sensor system for connecting to the Raspberry PI. Anything that makes prototyping faster, easier, and more efficient is good.


Comments: Another attempt to make attaching sensors a plug-n-play affair. If you plug this into arduino, there will still probably be a lot of code compilation to do.

Arduino Tre: The Agduino CPU


The Arduino Tre looks to be a great candidate as the CPU for agduino. According to the release notes, the Tre promises to deliver “100 times more performance with the Sitara-processor-based TRE than on the Arduino Leonardo or Uno. This performance opens the doors to more advanced Linux-powered applications.”

Other very useful features of the Tre:

  • Network via ethernet
  • MicroSD Slot
  • HDMI video port
  • Numerous analog and digital connectors

If we could build a .deb package that deploys simply and quickly: apt-get install agduino. Then the user just has to configure some .conf files, tell the duino where to find the sensors, servos, relays, etc., and it’s ready to deploy. It could be controlled via android, or simple web interface.

Check out Reef Angel for a good, but incomplete, analogy for this.

Agduino on G+

I’ve built a G+ community for people to interact specifically about topics related to arduino for agriculture. If you are interested in this topic, please join. There is a lot of potential to have some great interactions. G+ may or may not be the best platform. Let’s get people interacting and see where it goes.

Click on the link below to join:

G+ Agduino Community

Arduino Aquaponics in Oakland

Another arduino for aquaponics prototype. The economics of this remains of great interest. Sure the prototypes are expensive; there is a lot of development work going into it. But if the model is open source, development costs should go down exponentially once the platform matures. The step to jump start the process will be to deploy at greater scales. It will yield greater return on investment, and get sufficient leverage to accelerate the rate of progress.

You Don’t Need Pesticides

Spirochaete bacterium in compost

Image courtesy

An excerpt from my interview with Dr. Elaine Ingham. Thanks to Todd Hoff for the transcription.

Bacteria and fungi are really good at holding nutrients in their bodies. If you don’t have nutrients in the soil then you need bacteria and fungi. They also put up castle walls to protect plant root systems from insect pests and from fungal and bacterial diseases. The plant itself puts out certain kinds of sugars and proteins and carbohydrates to build that castle wall. By exuding different nutrients it attracts different bacteria and fungi to protect against particular diseases or hold certain nutrients. The plant has a lot of communication and control of what goes around its surfaces. Plants know how to protect themselves. You don’t need pesticides.