How we hackathon about the Internet of things did
Hi, Geektimes!
Most recently, we — Mail.Ru Group, Intel, and Unwired Devices — were doing a hackathon on the Internet of things. In principle, this could be an ordinary event - IoT hackathons are not just doing lazy right now. But we decided to come up with a format that would favorably distinguish us from other similar events.
And if we discard excessive modesty, then we can say that with this hackathon we are starting a new series of IoT events - events, on the basis of which we want to form a platform for discussing modern IoT technologies, which will be attended by not only technology enthusiasts , industry representatives and investors interested in the subject of the Internet of Things.
')
The insignificant part of coffee drunk in two days
We started with the hackathon. And now we want to tell you why we did it, what happened and what conclusions we drew.
What did we dislike about typical IoT hackathons? First of all, the fact that they, as a rule, do not have IoT as such - in the form in which it is understood in industry, utilities, or even a smart home. Unfortunately, in the educational IoT (and hakatons can be attributed to the educational segment) vacuum is now prevailing. If Z-Wave and ZigBee live and live in a smart house, the industry is looking at 6LoWPAN with interest, and the utilities plan to move to LoRa, then on an average hackathon you will most likely be offered to write something for a piece of hardware with regular Wi-Fi. In an Arduino development environment or similar. With analytics and visualization - and this will be the only serious part of such an event - in the cloud ThingWorx, Azure or the like.
Yes, even on such platforms, you can work out various software solutions and make sketches of IoT services, but generally speaking, you can also work out the same thing on a regular laptop. If you need to show how you can send data from an electric meter to the cloud, then Raspberry Pi and a laptop are equally suitable. And if you need to send this data from each of the 236 electricity meters of a real apartment building, both the Raspberry Pi and the laptop will not work equally well for this. That is, there is a quantitative difference between a laptop and RPi in such a task, but not a qualitative one.
We decided to change this and organize a hackathon on those IoT technologies that are not just used in industry today , but will be used tomorrow . Both hardware and software.
The second problem of many hackathons is closed, hermetic events, in which teams receive certain sets and instructions at the entrance, and prizes at the exit. Often nothing else happens: neither before, nor after, nor during (except, of course, the teams writing code). We also decided to change this - and, as I said above, turn the hackathon into a platform where there are teams, developers of the underlying technologies, industry representatives, and investors interested in IoT topics.
In terms of technology, the Internet of things falls into three areas:
Of these three subsystems, we decided to provide the hackathon participants ... of course, all three!
In fact, it is quite possible to build a wireless SCADA-system of an object on such components - from an office building to a factory site (wireless SCADA is of particular interest in the open spaces: if you can still build a small point in the building, and drive along it good old RS-485 + Modbus, it is simply impossible in some quarry, but at some large substation or water purification station - although it is possible, it is completely inhuman for the money).
It was interesting to look at the example of the list of participants: how the technologies used affect the content of the hackathon: by the beginning of the event, three main categories were clearly formed.
The first one is classic: teams, including student ones, that have their own project, and they would like to try to develop it within the framework of the hackathon. However, here the list of available equipment and software has already had an impact: some of the teams directly asked to bring them certain modules, for example, GPS or accelerometer.
The second is professional developers who were interested to stretch their hands, and at the same time to practice with things new to them, be it with MQTT and “wireless” or with Tarantool.
The third is teams from the business, not necessarily somehow related to IT, but interested in the technologies that we gave to the hackathon participants. For example, there was a team from a successful agricultural enterprise engaged in year-round cultivation of greenery. By their own admission, they were interested not so much in the process of creating something in two days, as in itself personal acquaintance with modern technology. A year-round agricultural enterprise in our latitudes is troublesome, requiring continuous monitoring of climate, watering and lighting (for me, for example, it was news that some plants have daily cycles, and shining the same lumen on them 24 hours a day is not the best the way to grow them), and many such farms are now eyeing the technologies of wireless monitoring and automatic control of greenhouses.
In general, an interesting mix came out: teams from the first group advanced their projects, teams from the second showed first how to work professionally and purposefully, and the third was about what we and some of the investors had to talk about - what technologies are interesting with in terms of potential customers.
We did not purposefully invite representatives of the industry to the first hackathon, although as a result they came, and not only as part of teams.
In the future, it is extremely interesting for us to develop cooperation with potential customers of what different teams can do on the hackathon - up to the format in which the hackathon is conducted under the guidance of one large customer who offers teams their business problems that need to be addressed as input. The advantages for customers here are obvious, but for teams it is a great opportunity to talk with people representing real business and knowing about its real problems, and not only those that the team is going to solve, but also those who are in the way of this solution.
In fact, such information is often lacking even for well-formed (but not yet released to a large market) start-ups. Even with a well-developed technological basis, they often do not realize how much economic, legal and political problems they can expect ahead. And for us it would be extremely interesting to give them feedback from the business, and this is a genuine feedback, sometimes even evil, and not a formal “well done, great project, work further”.
Here, of course, there is a dangerous gap between the reality of the business and the reality of the hackathon: the first expects to see the finished solution in a beautiful box and with a delivery of 10 working days from the moment of payment, the second shows a wiring and crutches construction made in two days and one night constantly something falls off. But as practice shows, if you approach the business wisely - and business can be taught to understand what startups are doing.
If we talk about investors, then, of course, on the one hand, the hackathon projects are not venture capital material, except perhaps some randomly misunderstood angel will take it under the wing. On the other hand, investors are interested in the state of things in the industry itself, it is interesting for them to see and hear what the beginning teams think in principle, and sometimes to follow their fate.
For example, our hackathon, although it took place after hours - the finale was on Sunday evening - was visited by representatives of CommIT Capital, IIDF, Skolkovo and others. Interviewed with many, contacts of some teams asked them to transfer.
Following the two-day work, each team made a 6-minute report on its development - and after two hours of reports (two dozen teams plus coffee breaks, so that everyone would have time to talk with the newly reported team in more detail) the jury left for the meeting.
We gathered the jury from representatives of all three organizers - and immediately abandoned the system of simple scoring, when each member of the jury puts something to each team, and then the scores are summed up. The main objection against such a system was that one member of the jury would stand for artistry, the other for technique - and the average temperature in the hospital would turn out.
This was fully confirmed: from the very first minute of discussing the draft, the jury was divided into three parts. Representatives of the Mail.Ru Group outweighed the technical side of the project, representatives of Intel - the commercial, well, we were somewhere in the middle, although with a penchant for the commercial side. As a result, instead of a formal dry vote, there was a lively discussion: if we exclude unequivocal outsiders unanimously, then the leaders, and even more so the distribution of seats between them changed during the discussion several times.
What we considered:
Actually, the main discussion was about finding a balance between the first and third points. On the one hand, a hackathon is a contest for writing code; on the other hand, we, firstly, arrange something more than the usual hackathon, and secondly, there are professional teams among the participants - and if you take into account only the written code, you can immediately give all the prizes to them.
In addition (and this is in contrast to the mechanized system of summation of points), a failure on any of the points could knock the team out of the number of finalists. For example, even the most beautiful presentation could not save a team that did not use Tarantool or, for example, spent the lion’s share of time discussing business models instead of writing code.
The first place is the “Agrogenesis” project: a monitoring system for agriculture (no, not the one that was mentioned above, that was called “GoodFarm” - IoT in agriculture is becoming very popular in general). The team quickly got on the list of winners and split the jury in two when trying to agree on its place in it: on the one hand, the existing team and project are visible, which she worked on before the hackathon and clearly continue after, on the other - not much was done on the hackathon a lot, and often needed the help of developers from our side.
The second place is the “Cool-IoT” team with a project of wireless sensors for the gym: the simulators (up to the dumbbells) are equipped with accelerometers and other sensors, which allow collecting statistics on the number of approaches, weight, etc. These systems are equipped with modern simulators, but they are very expensive, and many sports centers simply can not afford. The team for two days made a prototype of a dumbbell with an accelerometer (in the absence of iron - in the form of a plastic box), which really dumped data on the number of lifts of this dumbbell on 6LoWPAN in Tarantool, and the latter even displayed them in the web interface. The presentation team summed up a bit: there was little said about the possible implementation and commercialization details, and therefore some jury members doubted that such a project was technically and commercially possible.
The blue box is a dumbbell, although it is not very similar.
The third place is the “ColorControl” project, a smart lamp, the color of the illumination of which changes depending on the words pronounced alongside: say “deep ocean” - the lamp will gradually change color to blue, say “Sahara desert” to yellow. The project did not use the 6LoWPAN network: speech was processed on Intel Edison with recognition through the Google API. The team had an excellent presentation with a story about what kind of atmosphere such a lamp is capable of creating, for example, when reading a fairy tale to children at night, but a little let down by the use of exclusively third-party cloud services for basic data processing and incomprehensible prospects for commercialization: this lamp is interesting as a separate product, but it is unlikely to build a business around it.
In general, as you can see, choosing the winners was really difficult. Even now, when I write project descriptions, thoughts are beginning to creep in that, after all, another distribution of places could also be convincingly substantiated. In general, it would be our will, I think, the entire jury would agree simply to give 5-7 best projects in the first place to everyone.
The main conclusion from the hackathon: the next time, more preparation is needed from our side. In fact, the teams got two hardware and three software platforms at their disposal, which are interconnected with each other, and many of them have no experience with any of them. Therefore, it was often difficult even from the first approach to understand exactly which part of the whole system the problem got out of: in 6LoWPAN modules, in our Edison software, in Yocto Linux, in Tarantool, or how it all works on the night written from Saturday on sunday code. In practice, there were all the options, as well as their combinations.
Prior to the presentation for the organizers it was a mystery what it is and where it goes. It turned out - looking for leaks of harmful gases
But, by the way, this experience was also useful for us: it became clearer what questions (and not only technical ones) were often asked by inexperienced users, what they expected and what other documentation and examples should be prepared the next time.
If we speak thesis:
But these are technical issues. Speaking in general, we both liked the participants and we hope to further develop the theme of such events, gradually erasing the line between hackathons, start-up contests and just a thematic platform for discussing the development and needs of IoT.
We hope that this Industrial IoT hackathon (yes, we didn’t have a single smart home or a pedometer bracelet, but there were several teams with projects in agriculture at once!) - the first, but not the last. We are interested in the popularization of IoT, in the demonstration of the modern elemental, network and software base - for everyone who is interested in this just as we are.
But it is not very interesting for us to conduct a hackathon for the sake of a hackathon, therefore we are looking for partners among the business. If your company is interested in becoming the organizer of such a hackathon - to promote its products, to solve its problems or with any other good goals - we are ready to discuss a possible partnership with pleasure.
Contact us.
With this half-chair, Master Gambs begins a new batch of furniture ...
I. Ilf, E. Petrov, "The Twelve Chairs"
Most recently, we — Mail.Ru Group, Intel, and Unwired Devices — were doing a hackathon on the Internet of things. In principle, this could be an ordinary event - IoT hackathons are not just doing lazy right now. But we decided to come up with a format that would favorably distinguish us from other similar events.
And if we discard excessive modesty, then we can say that with this hackathon we are starting a new series of IoT events - events, on the basis of which we want to form a platform for discussing modern IoT technologies, which will be attended by not only technology enthusiasts , industry representatives and investors interested in the subject of the Internet of Things.
')
The insignificant part of coffee drunk in two days
We started with the hackathon. And now we want to tell you why we did it, what happened and what conclusions we drew.
What did we dislike about typical IoT hackathons? First of all, the fact that they, as a rule, do not have IoT as such - in the form in which it is understood in industry, utilities, or even a smart home. Unfortunately, in the educational IoT (and hakatons can be attributed to the educational segment) vacuum is now prevailing. If Z-Wave and ZigBee live and live in a smart house, the industry is looking at 6LoWPAN with interest, and the utilities plan to move to LoRa, then on an average hackathon you will most likely be offered to write something for a piece of hardware with regular Wi-Fi. In an Arduino development environment or similar. With analytics and visualization - and this will be the only serious part of such an event - in the cloud ThingWorx, Azure or the like.
Yes, even on such platforms, you can work out various software solutions and make sketches of IoT services, but generally speaking, you can also work out the same thing on a regular laptop. If you need to show how you can send data from an electric meter to the cloud, then Raspberry Pi and a laptop are equally suitable. And if you need to send this data from each of the 236 electricity meters of a real apartment building, both the Raspberry Pi and the laptop will not work equally well for this. That is, there is a quantitative difference between a laptop and RPi in such a task, but not a qualitative one.
We decided to change this and organize a hackathon on those IoT technologies that are not just used in industry today , but will be used tomorrow . Both hardware and software.
The second problem of many hackathons is closed, hermetic events, in which teams receive certain sets and instructions at the entrance, and prizes at the exit. Often nothing else happens: neither before, nor after, nor during (except, of course, the teams writing code). We also decided to change this - and, as I said above, turn the hackathon into a platform where there are teams, developers of the underlying technologies, industry representatives, and investors interested in IoT topics.
Technology
In terms of technology, the Internet of things falls into three areas:
- transmit data: wireless communication technologies that allow you to connect sensors and control systems;
- collect data: software and hardware system for collecting data from connected systems and storing them;
- process data: software environment for analyzing and visualizing accumulated data.
Of these three subsystems, we decided to provide the hackathon participants ... of course, all three!
- Transfer data: participants were available to build small mesh networks with 6LoWPAN protocol based on Unwired Kit modules - they included temperature and humidity sensors, soil moisture, gyroaccelerometers, relays, additionally there were buttons, GPS and other modules. Seeedstudio wired sensors could be connected to the kits.
- Collect data: a gateway between the 6LoWPAN network and Wi-Fi based on the Intel Edison microcomputer, which we specifically made for the event (but also, of course, with the future: the Edison is very interesting when you need good computational power and a lot of memory in the limiting compact size ).
- To process data: DBMS Tarantool , which in this incarnation has evolved from a DBMS into an IoT platform. Tarantool has a built-in application server with Lua - and it allows you to write both data collection scripts (our system gives them to MQTT, for example), and the processing of this data itself, with subsequent transfer to higher-level systems.
In fact, it is quite possible to build a wireless SCADA-system of an object on such components - from an office building to a factory site (wireless SCADA is of particular interest in the open spaces: if you can still build a small point in the building, and drive along it good old RS-485 + Modbus, it is simply impossible in some quarry, but at some large substation or water purification station - although it is possible, it is completely inhuman for the money).
Members
It was interesting to look at the example of the list of participants: how the technologies used affect the content of the hackathon: by the beginning of the event, three main categories were clearly formed.
The first one is classic: teams, including student ones, that have their own project, and they would like to try to develop it within the framework of the hackathon. However, here the list of available equipment and software has already had an impact: some of the teams directly asked to bring them certain modules, for example, GPS or accelerometer.
The second is professional developers who were interested to stretch their hands, and at the same time to practice with things new to them, be it with MQTT and “wireless” or with Tarantool.
The third is teams from the business, not necessarily somehow related to IT, but interested in the technologies that we gave to the hackathon participants. For example, there was a team from a successful agricultural enterprise engaged in year-round cultivation of greenery. By their own admission, they were interested not so much in the process of creating something in two days, as in itself personal acquaintance with modern technology. A year-round agricultural enterprise in our latitudes is troublesome, requiring continuous monitoring of climate, watering and lighting (for me, for example, it was news that some plants have daily cycles, and shining the same lumen on them 24 hours a day is not the best the way to grow them), and many such farms are now eyeing the technologies of wireless monitoring and automatic control of greenhouses.
In general, an interesting mix came out: teams from the first group advanced their projects, teams from the second showed first how to work professionally and purposefully, and the third was about what we and some of the investors had to talk about - what technologies are interesting with in terms of potential customers.
Guests
We did not purposefully invite representatives of the industry to the first hackathon, although as a result they came, and not only as part of teams.
In the future, it is extremely interesting for us to develop cooperation with potential customers of what different teams can do on the hackathon - up to the format in which the hackathon is conducted under the guidance of one large customer who offers teams their business problems that need to be addressed as input. The advantages for customers here are obvious, but for teams it is a great opportunity to talk with people representing real business and knowing about its real problems, and not only those that the team is going to solve, but also those who are in the way of this solution.
In fact, such information is often lacking even for well-formed (but not yet released to a large market) start-ups. Even with a well-developed technological basis, they often do not realize how much economic, legal and political problems they can expect ahead. And for us it would be extremely interesting to give them feedback from the business, and this is a genuine feedback, sometimes even evil, and not a formal “well done, great project, work further”.
Here, of course, there is a dangerous gap between the reality of the business and the reality of the hackathon: the first expects to see the finished solution in a beautiful box and with a delivery of 10 working days from the moment of payment, the second shows a wiring and crutches construction made in two days and one night constantly something falls off. But as practice shows, if you approach the business wisely - and business can be taught to understand what startups are doing.
If we talk about investors, then, of course, on the one hand, the hackathon projects are not venture capital material, except perhaps some randomly misunderstood angel will take it under the wing. On the other hand, investors are interested in the state of things in the industry itself, it is interesting for them to see and hear what the beginning teams think in principle, and sometimes to follow their fate.
For example, our hackathon, although it took place after hours - the finale was on Sunday evening - was visited by representatives of CommIT Capital, IIDF, Skolkovo and others. Interviewed with many, contacts of some teams asked them to transfer.
Jury and Judging
Following the two-day work, each team made a 6-minute report on its development - and after two hours of reports (two dozen teams plus coffee breaks, so that everyone would have time to talk with the newly reported team in more detail) the jury left for the meeting.
We gathered the jury from representatives of all three organizers - and immediately abandoned the system of simple scoring, when each member of the jury puts something to each team, and then the scores are summed up. The main objection against such a system was that one member of the jury would stand for artistry, the other for technique - and the average temperature in the hospital would turn out.
This was fully confirmed: from the very first minute of discussing the draft, the jury was divided into three parts. Representatives of the Mail.Ru Group outweighed the technical side of the project, representatives of Intel - the commercial, well, we were somewhere in the middle, although with a penchant for the commercial side. As a result, instead of a formal dry vote, there was a lively discussion: if we exclude unequivocal outsiders unanimously, then the leaders, and even more so the distribution of seats between them changed during the discussion several times.
What we considered:
- Work on the hackathon. Since about half of the members of the jury during the hackathon were engaged in continuous consulting of teams on technical issues, and some did not even leave the office for the night, we knew very well which team had a lot of time and what it did, and which part of it did it on its own.
- Using the proposed technology. At a minimum, Tarantool should have been used without fail.
- Commercial component. As far as the project (or at least its idea) is interesting from the point of view of embodiment in reality.
- The quality of the presentation. Of course, professional performance was not required, technical problems with the demonstration of projects (and they were about everyone who tried to demonstrate something in the work) also did not pay attention - but, at least, it was necessary to explain what the project is and why He is needed in this world so that all members of the jury understand this.
Actually, the main discussion was about finding a balance between the first and third points. On the one hand, a hackathon is a contest for writing code; on the other hand, we, firstly, arrange something more than the usual hackathon, and secondly, there are professional teams among the participants - and if you take into account only the written code, you can immediately give all the prizes to them.
In addition (and this is in contrast to the mechanized system of summation of points), a failure on any of the points could knock the team out of the number of finalists. For example, even the most beautiful presentation could not save a team that did not use Tarantool or, for example, spent the lion’s share of time discussing business models instead of writing code.
Winners
The first place is the “Agrogenesis” project: a monitoring system for agriculture (no, not the one that was mentioned above, that was called “GoodFarm” - IoT in agriculture is becoming very popular in general). The team quickly got on the list of winners and split the jury in two when trying to agree on its place in it: on the one hand, the existing team and project are visible, which she worked on before the hackathon and clearly continue after, on the other - not much was done on the hackathon a lot, and often needed the help of developers from our side.
The second place is the “Cool-IoT” team with a project of wireless sensors for the gym: the simulators (up to the dumbbells) are equipped with accelerometers and other sensors, which allow collecting statistics on the number of approaches, weight, etc. These systems are equipped with modern simulators, but they are very expensive, and many sports centers simply can not afford. The team for two days made a prototype of a dumbbell with an accelerometer (in the absence of iron - in the form of a plastic box), which really dumped data on the number of lifts of this dumbbell on 6LoWPAN in Tarantool, and the latter even displayed them in the web interface. The presentation team summed up a bit: there was little said about the possible implementation and commercialization details, and therefore some jury members doubted that such a project was technically and commercially possible.
The blue box is a dumbbell, although it is not very similar.
The third place is the “ColorControl” project, a smart lamp, the color of the illumination of which changes depending on the words pronounced alongside: say “deep ocean” - the lamp will gradually change color to blue, say “Sahara desert” to yellow. The project did not use the 6LoWPAN network: speech was processed on Intel Edison with recognition through the Google API. The team had an excellent presentation with a story about what kind of atmosphere such a lamp is capable of creating, for example, when reading a fairy tale to children at night, but a little let down by the use of exclusively third-party cloud services for basic data processing and incomprehensible prospects for commercialization: this lamp is interesting as a separate product, but it is unlikely to build a business around it.
In general, as you can see, choosing the winners was really difficult. Even now, when I write project descriptions, thoughts are beginning to creep in that, after all, another distribution of places could also be convincingly substantiated. In general, it would be our will, I think, the entire jury would agree simply to give 5-7 best projects in the first place to everyone.
Experience son of difficult mistakes
The main conclusion from the hackathon: the next time, more preparation is needed from our side. In fact, the teams got two hardware and three software platforms at their disposal, which are interconnected with each other, and many of them have no experience with any of them. Therefore, it was often difficult even from the first approach to understand exactly which part of the whole system the problem got out of: in 6LoWPAN modules, in our Edison software, in Yocto Linux, in Tarantool, or how it all works on the night written from Saturday on sunday code. In practice, there were all the options, as well as their combinations.
Prior to the presentation for the organizers it was a mystery what it is and where it goes. It turned out - looking for leaks of harmful gases
But, by the way, this experience was also useful for us: it became clearer what questions (and not only technical ones) were often asked by inexperienced users, what they expected and what other documentation and examples should be prepared the next time.
If we speak thesis:
- We will try to more popularize the capabilities of Tarantool and Lua - the hackathon participants were interested primarily in specific examples of projects and opportunities;
- We will pay more attention to the documentation of the “Start started” class - both for hardware and software;
- The next version of the Unwired Kit will be fully plug'n'play - now the radio modules of the firmware are not universal, they must sometimes be changed depending on the connected expansion cards.
But these are technical issues. Speaking in general, we both liked the participants and we hope to further develop the theme of such events, gradually erasing the line between hackathons, start-up contests and just a thematic platform for discussing the development and needs of IoT.
What's next?
We hope that this Industrial IoT hackathon (yes, we didn’t have a single smart home or a pedometer bracelet, but there were several teams with projects in agriculture at once!) - the first, but not the last. We are interested in the popularization of IoT, in the demonstration of the modern elemental, network and software base - for everyone who is interested in this just as we are.
But it is not very interesting for us to conduct a hackathon for the sake of a hackathon, therefore we are looking for partners among the business. If your company is interested in becoming the organizer of such a hackathon - to promote its products, to solve its problems or with any other good goals - we are ready to discuss a possible partnership with pleasure.
Contact us.
Source: https://habr.com/ru/post/396807/
All Articles