How to create a peer-to-peer communication over RF among ARA-based devices?

[Peter Babel]

I'm putting this out there for you, developers, who might be interested in working with us on this project...

In our implementation of the ARA platform, we'd need anywhere from 2 - 1000s of ARA devices to communicate with each other via a "closed-circuit" RF protocol.

In the beginning, we would definitely want to start with the ability to manage the communication with at least 10-500 devices.

Components:

- ARA RF module that RECEIVES the signal from the other devices on the closed-circuit network (read: host/master)

- ARA RF module that SENDS the signal to the RECEIVER (read: participant)

  - Could these two be one and the same module, but with different 'access rights' to the 'session'?

  - What RF tech should be used if the application is for standard hotel/convention venues with large population of Wi-Fi and/or other RF interference AND rather long range (cca 600ft) requirement?

In some cases (at some events), the closed-circuit RF network could be complemented by a cellular/Wi-Fi one at the same time, so that not only the participants INSIDE the venue can participate in the communication, but also people on the Internet.

- It might be sufficient if Wi-Fi/cellular ARA module is only implemented in the RECEIVER (master) device so it can connect to the Internet-based session and allow the outsider to be considered part of the group.

- All other devices (SENDER/participant) only need to be connected to the master, so they would only need the "closed-circuit" ARA module installed

This variability would allow the MASTER device to have a smaller battery but more capable, feature-loaded components (while needing to be connected to permanent source of power - A/C adapter, which is OK in this usage scenario). At the same time, the PARTICIPANT devices would need more battery life and would only need one type of specific feature - RF module.

We're looking for intellectual input and partners in making the actual RF module when the ARA platform is released to the development community.

[Elender Góis Gallas]

I have liked of the idea, but I think that using RF modules is not the best way to do it. If is a peer-to-peer communication so just Wi-fo would do the job.

[Peter Babel]

In this application, we need the peer-to-peer network to be secure and 'closed' to only those devices that know how to connect to each other.

Also, I don't want to have to set up another wireless network at a venue that already might have 10+ Wi-Fi networks running...

[Mike Trieu]

That's what the 802.11s standard already describes. Peers authenticate via the Authenticated Mesh Peering Exchange (AMPE).

[Peter Babel]

I'd like to work with someone on integrating 802.11s radio into an ARA module. Any takers?

[Jason Wright]

are we describing citizens band telephony here? so a module would act as a sender, a receiver, and a repeater, creating a unified grid constructed exclusively from ara devices within range of each other. it would work well in urban environments. it needs world class encryption to keep it secure. carriers would dislike it. governments would dislike it. for those two reasons alone it needs to happen.

[Matei-Eugen VASILE]

I second this.

WiFi-based mesh networks exist now too, but they haven't been gaining much traction among the wide public, most likely because the end-users want something that work out-of-the-box, and that hasn't been done yet. Maybe and Ara module that provides out-of-the-box mesh networking functionality would be what brings it into the mainstream. Count me in as interested in this.

[Charles Dylan Reilly]

The idea that Ara Phone users create a mesh wi-fi network, connecting with each other and some to the internet blows mind. Count me in on this project

Sent from my iPhone

--
You received this message because you are subscribed to the Google Groups "Ara Module Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ara-module-devel...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ara-module-developers/fd31e975-9f9f-47c6-b94a-e05757ad34b7%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Peter Babel]

Loving the energy on this already, but I want to make something clear now...

We're a legit company dealing with interactive devices for meetings and events RIGHT NOW and I'm trying to find a creative solution to the problem of distributing 100s to 1000s of 'keypads' to the attendees of these events when the devices, much like the events themselves are different from one to the other.

Modular design would allow us to have a base/endo with multiple possible configurations...where the center point (for the purposes of having feedback collection from a group of people) is a RF chip that I'd like to source here, in ARA Community.

Otherwise, I'm open to all non-anarchic :) ideas, LOL.

P.

[Matei-Eugen VASILE]

The idea is clear: mesh networking Ara module. How people are going to use it afterwards is their own business. The technology itself is purpose-agnostic.

[Cyberteknix]

Hello!

Remember, the lower the frequency, the more "permeable" the RF will be at cost of bandwidth.  Anything greater than 2.4GHZ could bounce around quite a bit.  However with the advent of 802.11AC and MIMO+beamforming, things start to get extremely exciting:  http://en.wikipedia.org/wiki/IEEE_802.11ac

Secondly, (I may be WAY off base) here are a couple of options that have low power and low cost requirements.

There are 433mhz RF chipsets that are low bandwidth (5kbps), yet more "near field" than anything.  Upping the wattage, and a low pass filter could produce decent results for your keypads at extremely low cost.

Here is a (5kbps) 433mhz concept:

http://www.seeedstudio.com/depot/433Mhz-Wireless-Serial-Transceiver-Module-1-Kilometer-p-1733.html

Here is a 900mhz (50kpbs) concept:

http://www.freaklabs.org/index.php/Blog/Chibi/Introducing-the-Freakduino-900-MHz-Long-Range-Wireless-Board-v2.1a.html

Now the only other issue is aggregation. To manage 1000 simultaneous connections would require a pretty complex array, but very possible.  In my estimation it would require multiple (supporting ~20 clients per node) "access points" which then work like cell towers.  If a device comes online/in range, it takes affinity to a particular tower.  Heartbeat triangulation would have to be employed here, and you could feasibly detect how people are moving around the event while accessing specific "nodes/towers" .. 

A little off-topic, but my goal for these is to integrate them into automation for distal sensor management of "robotic" farming equipment.  My usage would primarily be "simplex" as they would just "report in" to a central base.

[mike]

There is a group of us working on a wireless mesh routers which operate at the discretion of the user with a new cryptocurrency in development called Skycoin, https://bitcointalk.org/index.php?topic=380441.0 . I attended the conference since we may make a wireless mesh networking module for the Ara as well. Quantenna claims 12-21 dB gain with their 802.11ac chip, the QSR1000 series with 4 streams. I haven't been able to find a white paper or data sheet for this. Broadcom has BCM4360 with 3 streams, the http://www.broadcom.com/products/Wireless-LAN/802.11-Wireless-LAN-Solutions/BCM4360 . I also met a chip and module manufacturer for TV whitespace networking, with 12 mbps data rate at a range of 10s of kilometers. This would enable units to fallback to units at long range if there is nothing at close, high speed 802.11ac range, maybe .1-1 km. I'm looking at software defined radio and fractal antenna arrays to enable use of the same antennas for a wider range of frequencies and reduce their size. I will begin a new thread for those wishing to discuss wireless mesh networking for the Ara.

[Peter Emmenegger]

Peter - have a look at http://mmbnetworks.com/modules . Their offering may work really nicely with a good chunk of work already done for you.

On Tuesday, April 15, 2014 6:21:43 PM UTC-4, Peter Babel wrote:

[Jeff Drake]

I would definitely be interested in working on integrating 802.11s, or another meshing technology into an ARA module.  

Jeff Drake

[Peter Babel]

Thank you, everyone for your ideas. I am monitoring the Dev Forums daily, so if anyone wants to get in touch and start developing the hardware or software when the MDK and tools are available, I'm here, waiting :)

P.

[Daniel Beckstein]

My ideas (Theoretic concept)
Assuming the communication is mostly text, not images:

Most RF modules usually are able to be both (emitter, receiver), please correct me if I am wrong.
-> it's not possible to separate reciever/emitter on hardware/modules

Problem: Different session access rights
Solution:
    - Impose some kind of data limit for each device,
    - use different encryption keys for each user, and one "semi public" key (known by your devices),
      for simultanious broadcasting to your devices

Problem: Interference
Solution:

 - Allocate 20 time slots for your devices (0.25s each -> 20x0.25 = 5s latency/lag) Final lag -> 5s+1s (1s for the broadcasting message to all users)
 
   Devices in Group 1 send/or recieve the first 0.25s of every 5 sec intervals,
   Devices in Group 2 send/or recieve the first 0.25s of every 5 sec intervals...
  
   Should enable some kind of scalable network (trade off between latency, bandwidth vs number of users)
  
   Number of Users: 2000 -> 100 user in each time slot (eventually asing groups of 10 users to mircotimeslots)
              
   Note: In this example you just can use "0.25*device bandwidth/100" = 360bit/6s  per user for upload
  
            360bit = 44 Byte -> 30 Byte -> >60 characters (with compression applied) which should be well suited for short text messages (try typing 60characters/6sec)
           
   This system should be able to deliver 18KB/6s to all users with the broadcast Message each 6 s,
   so the total lag during the event should be < 12 seconds  (from posting a message by single user - till everybody has this message delivered)

    If 12 sec is too high, just use this system for 1000 instead of 2000 users and then it should be ~7s

   Calculated with:   
   Example RF module:  XBee - https://www.sparkfun.com/products/9097
   600mW, 156kbit/s, 40$  (power usage could be too high?)
       
 - Avoid 2.4 GHz, by using 900 mHz (which also enables a higher range)
   I don't know how crammed the 900 mHz range usually is, but at least you won't interfere with WIFI and LTE (<800mHz, >1.900mHz)
  
Use a Master Server <300$ (AMD Kabini on ITX-Board with Laptop-AC Adapter; +8GB RAM; 35 watt; http://www.newegg.com/Product/Product.aspx?Item=N82E16813157491;),
which handles encryption, the key management, message routing to the web and to the RF network, decompression, compression)

This Master Server should allow all your RF participants to access the posts of the external members too (which are connected via web)

Keep bandwith low:
- use compression on every message
- maybe limit characters to UTF-8

If I have time, I will further think about dedicated timing of broadcasting, and secret messaging to single users and write it more structured

Images: - Description of system
             - Message Forwarding

[Peter Babel]

Daniel,

thank you for your contribution. I think I'll need your input once the development of the actual module can start.

Would you be interested in working on this with me and whoever else wants to participate?

[Mehdi Zakaria Benadel]

As people already stated it, mass objects wireless is covered by ZigBee (or IEEE 802.15.4). But it will not be the only standard applying this in a few years, because LTE B is planning on integrating data transfer between mobiles, and Bluetooth Low Energy is also beginning to spread.

You should first choose the right one for you before going into any development, given that they have their own limitations (ZigBee is really limited in bitrate), different implementations regarding the countries, and technologically, some might be simpler and other might be a better insvestment in the long term (for example, WiMax has been wiped off by LTE success, and WiFi may follow in a few years).

[Mark Bastiaans]

Short version: I would avoid putting mesh networking in a specific Ara module unless you want to heavily optimize for a specific use case or maturity of implementation.

Long version: for mesh networking, you need to implement a couple of OSI layers some way or another, either in hardware or in software. An integrated stack (i.e. a stovepipe) usually sacrifices a degree of freedom for things like power consumption, performance and scalability. Zigbee can be considered to be such a stack and I think it will be relatively easy to make an Ara Zigbee module.

You can also choose for a generic radio / specific routing protocol approach, which will give you more freedom in how you implement and use your low-level OSI layers at the possible (note: possible) expense of power consumption, performance and scalability. Which looks a lot like the tradoff that Project Ara makes.

We're currently using off-the-shelf WiFi sticks and a userspace ad-hoc routing protocol (OLSRv2 packaged in a Docker container) for mesh networking tens of Raspberry Pis. Theoretically you can use any radio with this routing protocol as long as its link state is exposed to the protocol (in our case, the WiFi chipset and its Linux kernel drivers and firmware support it). For example, we've briefly looked at OLSRv2 on Bluetooth LE. For our use case, which is still in the prototyping stage, the freedom of experimenting with different routing protocols (i.e. making them modular in software) and being able to virtualize network functions (routing protocols, but also security stuff) outweighs all possible disadvantages of this approach.

[Aurelio Gouvêa de Melo]

Returning to the idea of ​​creating a technology for connecting pear-to-pear among users, I believe that Wifi consumes more energy than the 2G, would not be interressante if each user work as client and server simultaneously?

If each user manage the connection for the next 16 or 32 users, the number of data that it will take care of will be smaller. It can also store who are 256 or 1024 users on the next level to facilitate searches. Once you establish a route to the desired user it will start looking for alternatives to make the connection in case of movement between the parts and negotiate an encryption between them. As in general there will be multiple vias to conect the users the protocol can use this advantage to ensure an safe conection. 

We need to modify only the physical layer.

To unsubscribe from this group and stop receiving emails from it, send an email to ara-module-developers+unsub...@googlegroups.com.

[m d]

Do you have a link to your hardware and software configuration?

> --
> You received this message because you are subscribed to a topic in the Google Groups "Ara Module
> Developers" group.
> To unsubscribe from this topic, visit
> https://groups.google.com/d/topic/ara-module-developers/d6vQtxFaTNU/unsubscribe.
> To unsubscribe from this group and all its topics, send an email to
> ara-module-devel...@googlegroups.com
> <mailto:ara-module-devel...@googlegroups.com>.

> To view this discussion on the web visit

> https://groups.google.com/d/msgid/ara-module-developers/8213951c-a86a-4956-8119-159483896bc4%40googlegroups.com
> <https://groups.google.com/d/msgid/ara-module-developers/8213951c-a86a-4956-8119-159483896bc4%40googlegroups.com?utm_medium=email&utm_source=footer>.

> For more options, visit https://groups.google.com/d/optout.

--
to connect on Linkedin,
http://www.linkedin.com/inviteFromProfile?from=profile&key=15253674&firstName=Mike&lastName=Doty

[Mark Bastiaans]

We're currently finishing our project (which is about mass provisioning and configuring lots of embedded devices for use in mesh networks) and are open to releasing parts of our code as open source. As soon as a full write-up on our project and source code is available, I'll post the link here. But here's a short summary.

Hardware: Raspberry Pi model B, PiFace controller / LCD board, TP-Link TL-WN722N WiFi, Plugable USB Bluetooth 4.0 LE

Glue: Arch Linux ARM, Docker (packaging and lightweight virtualization), Puppet (mass provisioning and configuration), bleno (NodeJS Bluetooth 4 LE lib) and custom scripting

Test cases: three-stage provisioning (initial bootstrapping, device-specific configuration, user-specific including key exchange), OLSRv2 (we put some work in attaching physical network interfaces to Docker containers), TINQ (VPN in ad-hoc networks) and a custom mesh network chat app

Range indoor is about 20 metres through a couple of office walls, and that's without adjusting link parameters.

> ara-module-developers+unsub...@googlegroups.com
> <mailto:ara-module-developers+unsub...@googlegroups.com>.

[mike]

Thanks. Looks like you're designing for within a building or a plant. Look forward to seeing your write-up.

[99guspuppet]

The ideas in this post are excellent.
I am interested in a network centered on a human body .

Two of the ISM bands allowed in the USA are  
433.50MHz - 434.79MHz  ( span of 1.74 MHz )
902MHz - 928MHz   ( span of 26MHz )
ISM ==> Industrial-Scientific-Medical

I favor 433 ISM band because it is low cost... has good penetration of walls.


=====    TI CC1101   =========
I am using the TI CC1101 as an example... other chip makers do similar things.

The CC1101 from TI can be purchased on a PCB for $4.12 on eBay.
TI provides design documents and design tools.
Here is a design document    http://www.ti.com/lit/sg/slya020a/slya020a.pdf
Here is a design tool    http://www.ti.com/lsds/ti/wireless_connectivity/tools_software.page#smartrfstudio

If we chose to use 12.5KHz channel spacing. Data rate would be perhaps a max of 2500 bits per second.  Range could be 2000 meters in free space.
In the 433 ISM band we could have as many as 139 channels.    ( 1.74 x 10**6 ) divided by ( 1.25 x 10**4 )
In the 900 ISM band we could have as many as 2080 channels.   ( 2.6 x 10**7 ) divided by ( 1.25 x 10**4 )
  

[Garrett Kinsman]

This is fascinating. Take a look at opengarden.com they're building WiFi based mesh networks that work right now. Most of the technology is in our devices already, what lacks is a the software. (4.4 removed much of this functionally unfortunately)

The democratising of these antennas will allow for extremely complex networks. Its totally possible. Who will be the first to do it?

[Garrett Kinsman]

Guys I've been infatuated with this idea since I first heard about Ara. I know the people that can do this from a hardware/ software perspective. I'd seriously like consider putting some sort of concepts together. I'm a graphic designer and can create mockuops/ ui. We could probably convince the right people if we gather the right knowledge, people, and pretty pictures to impress investors.

I just moved to the bay area, and would like to know if any of you are nearby... I would like to talk more. my site: gekinsman.blogspot.com

[M Doty]

Those are some really cool concepts on your site. We have a meetup meeting every other wednesday, usually in Millbrae, http://www.meetup.com/Project-Ara-Developers-SF-Bay-Area/events/187445012/ . I may have to postpone this weeks meeting since I'm out of town, and it's soon after the holiday anyway. A group of us are working on some modules, and we can use more team members. The plan is at https://docs.google.com/document/d/1godyapfJdEmchmsuJYBntJMu3kHXJwTXMqdnX1lCZ4k/edit?usp=sharing . Please let me know if either of you are interested in working on the projects.

On Sat, Jul 5, 2014 at 11:14 PM, Garrett Kinsman <bossm...@gmail.com> wrote:

Guys I've been infatuated with this idea since I first heard about Ara. I know the people that can do this from a hardware/ software perspective. I'd seriously like consider putting some sort of concepts together. I'm a graphic designer and can create mockuops/ ui. We could probably convince the right people if we gather the right knowledge, people, and pretty pictures to impress investors.

 I just moved to the bay area, and would like to know if any of you are nearby... I would like to talk more. my site: gekinsman.blogspot.com

--
You received this message because you are subscribed to a topic in the Google Groups "Ara Module Developers" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/ara-module-developers/d6vQtxFaTNU/unsubscribe.

To unsubscribe from this group and all its topics, send an email to ara-module-devel...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ara-module-developers/bd5ca7a4-87dc-4e23-aebc-3bade444ef28%40googlegroups.com.

[Matthew J Wronkiewicz]

You might also want to take a look at the DTN implementations for Android, especially IBR-DTN.

DTN is designed to handle packet delivery in frequently-disconnected networks like remote villages where the only internet access is an Android phone on a mule. Or on a Mars rover that has no line of sight to Earth.

The IBR-DTN Android node <http://goo.gl/NH1Rz9> already supports WiFi-Direct. The Linux node also includes convergence layers for 802.15.4 and other networks.

The downside is that it's not TCP/IP. Any applications you want to run on it will have to be written to account for its UDP-like qualities and potentially week-long latencies. Like NASA is doing right now (in the most roundabout way possible) with astronaut email on the ISS <http://www.topcoder.com/dtn/astronaut-email/>.

Matt

[Garrett Kinsman]

I would definitely like to attend. The work you guys are doing will change the world. I think I might be able to connect you guys with the people that have working software and funds. What stages of development are you guys in?

[Mikhail Paremski]

I love it. Here's how I mentally broke down the project - GPS, computing, communication, battery, membership

GPS: Since project ara is a personal mobile device, you can't model this mesh network as a bunch of static points that occasionally move. You have to work under the assumption that all of these devices are constantly moving, so first and foremost, you'll need to know their location with respect to one another. You want to have as few transmissions as possible, so you want devices to relay the message as close to their maximum radius of transmission as possible, and you'll need to compute these paths based on the GPS coordinates of the devices.

Computing: Going back to the issue of a bunch of devices constantly moving with respect to one another is that you inevitably have to do some computing to do path optimization to get some information from one peer, to another peer a few (or more) devices away. One way of keeping communication paths optimized is to have a server somewhere that does all of this computing, but that would make this mesh network a subscription service, which would in essence equate to a carrier, which is (at least for me) the primary appeal for a peer-to-peer network - no carrier. An alternative is using the mentality behind http://www.flashmobcomputing.org/. Essentially it ties a bunch of computers together, and makes them act as a single computer. I'm not sure how big a computing hit all this computing would be to each device, but if the module had a chip to contribute to this "flash mob mesh network computer", I would think membership in this mesh network would be more appealing - you're not taking a performance hit on the rest of your device by installing this module. In addition to processing power, shared "flash mob memory" would be an interesting concept to play with to store the member database, but more on that later, in the next section actually.

Communication: As people have mentioned there are a variety of technologies to create a peer to peer network. In my opinion, since we're talking about Project Ara here, it doesn't make sense to invest in any hardware for this module to make the peer to mesh network. It would be more prudent to have hardware requirements, with recommended modules. The logic here is that people will most likely want, or already have some "conventional" rx/tx module to begin with. Instead of integrating some wifi radio into this module, you can piggyback on any other wifi module. And if you REALLY want to get crazy, why limit the network to only one form of communication? Some people may have only Bluetooth on their device, some may have only WiFi, some may have both. What if someone with a Bluetooth only device wants to communicate with a WiFi only device? Send it through a device that transforms the Bluetooth communication to WiFi using the micro controller in this module. In fact, you can even keep various communication methods in mind during the process of optimizing a path for a packet - it may be faster, or more energy efficient to bounce a signal to a device with LTE or 2G/3G  to send it long distance then convert it back to the method that the receiving device is capable of receiving. Also, at some point devices must volunteer access to the internet, so devices with access to WiFi that goes out to the wired network will have to interface with the P2P module (which is another option for transmitting data long distances).

Battery: Every device essentially takes on the role (or part of the role) of a server that computes these optimized paths. This is naturally going to put a drain on the battery, so this module would have to either recommend additional battery capacity, or provide it in the module itself.

Membership: The thing about peer to peer networks is that they don't work without peers. While I'm sure there will come a time when "P2P network module" will become commonplace to the point where people who are assembling their Ara-like device are thinking, "okay, first I need a P2P module, and maybe a WiFi or BT module to go with it" as opposed to, "Alright I need a WiFi module, oh hey, look at this really awesome P2P module that works with my WiFi module!", but I don't think we're there yet. I figure the way you can get public awareness there is by partnering with other module manufacturers, retail stores, or with Google. The way this would work is for example - you partner with a wifi module manufacturer, and they put in a sentence on the back of their product box that says, "works with P2P mesh network for wifi everywhere you go" (yeah I'm not a marketing guy, so I'm sure someone will come up with a catchier phrase). Or in the case of retail stores, perhaps you can work out an agreement where customers get a sweet deal on the P2P module if they buy an Ara base, or module. Last but not least, maybe you pitch your idea well enough to Google, where Google says, "This is awesome, everyone that buys an Ara with this module included automatically contributes to the mesh network, expands it, and makes it available to more and more people without the hassle of searching for free wifi, or paying for a carrier data plan, which means people can access the Google search engine and adverts more often and more conveniently, so we'll go ahead and prepackage this P2P module with every Ara base" If you get REALLY lucky, Google may just buy you outright, and include your technology in some more conventional phones =).

Anyhow, that's my rant. I'm really excited to see where this goes.

[Garrett Kinsman]

Imagine the possibilities for collaborative, mesh based computing. Also I'm thinking Satellites... Heres a graphic I did:

I'm pretty much convinced the internet will one day be completely free, and built on the backbone of our mobile devices.

[holly]

hello, Peter Babel.

i am holly.

i already pay attention to your project ara for a long time. our company's name is xi'an tianlong science and technology co., ltd.

we have a kind of portable ATP hygiene monitoring system.i hope we have chance to cooperate with you.

my email address:hollyl...@gmail.com.

Hope to hear from you soon.

在 2014年4月25日星期五 UTC+8上午3:42:15，Peter Babel写道：

[AMIT]

On Wednesday, April 16, 2014 at 12:04:02 PM UTC+5:45, Peter Babel wrote:

In this application, we need the peer-to-peer network to be secure and 'closed' to only those devices that know how to connect to each other.

Also, I don't want to have to set up another wireless network at a venue that already might have 10+ Wi-Fi networks running...

On Tuesday, April 15, 2014 9:19:50 PM UTC-4, Elender Góis Gallas wrote:

I have liked of the idea, but I think that using RF modules is not the best way to do it. If is a peer-to-peer communication so just Wi-fo would do the job.

we can try miniaturized male female wifi antenna (i.e beam antenna log- periodic antenna). They have way more range, and are more precise and flexible. 

[Garrett Kinsman]

Peter,

The solution is simple, at least in my opinion. The easiest solution is to use the existing BLE chipset already implemented in the Android device. 

Spiral 3 will have an included chipset that will far exceed anything (in terms of efficiency and cost) that could be manufactured in small volume for your specialized application.

The secret sauce that will make this all work is currently in Google's hands. Right now the parts of Android code that make devices connect P2P is a tad broken. Much of the problem lies in the fragmentation of Android- Ara should fix that. Those of us in the mesh community are hoping to see some updates at IO. 

[Peter Babel]

Now this is great news! I am excited for this year's I/O as I hear ARA team will be talking about their progress so far.

Your idea makes complete sense, but yet, we'll have to get Google to work on some of the glue...and other parts.

Thanks for the feedback and please stay in touch.

P.

[Cummings]

Sent from an iPhone, Awesome!

On Thursday, April 17, 2014 at 10:24:20 AM UTC-5, Charles Reilly wrote:

The idea that Ara Phone users create a mesh wi-fi network, connecting with each other and some to the internet blows mind. Count me in on this project

Sent from my iPhone

On Apr 17, 2014, at 3:37 AM, Jason Wright <jasonpa...@gmail.com> wrote:

are we describing citizens band telephony here? so a module would act as a sender, a receiver, and a repeater, creating a unified grid constructed exclusively from ara devices within range of each other. it would work well in urban environments. it needs world class encryption to keep it secure. carriers would dislike it. governments would dislike it. for those two reasons alone it needs to happen.

On Tuesday, April 15, 2014 11:21:43 PM UTC+1, Peter Babel wrote:

I'm putting this out there for you, developers, who might be interested in working with us on this project...

In our implementation of the ARA platform, we'd need anywhere from 2 - 1000s of ARA devices to communicate with each other via a "closed-circuit" RF protocol.

In the beginning, we would definitely want to start with the ability to manage the communication with at least 10-500 devices.

Components:

- ARA RF module that RECEIVES the signal from the other devices on the closed-circuit network (read: host/master)

- ARA RF module that SENDS the signal to the RECEIVER (read: participant)

  - Could these two be one and the same module, but with different 'access rights' to the 'session'?

  - What RF tech should be used if the application is for standard hotel/convention venues with large population of Wi-Fi and/or other RF interference AND rather long range (cca 600ft) requirement?

In some cases (at some events), the closed-circuit RF network could be complemented by a cellular/Wi-Fi one at the same time, so that not only the participants INSIDE the venue can participate in the communication, but also people on the Internet.

- It might be sufficient if Wi-Fi/cellular ARA module is only implemented in the RECEIVER (master) device so it can connect to the Internet-based session and allow the outsider to be considered part of the group.

- All other devices (SENDER/participant) only need to be connected to the master, so they would only need the "closed-circuit" ARA module installed

This variability would allow the MASTER device to have a smaller battery but more capable, feature-loaded components (while needing to be connected to permanent source of power - A/C adapter, which is OK in this usage scenario). At the same time, the PARTICIPANT devices would need more battery life and would only need one type of specific feature - RF module.

We're looking for intellectual input and partners in making the actual RF module when the ARA platform is released to the development community.

--

You received this message because you are subscribed to the Google Groups "Ara Module Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ara-module-developers+unsub...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/ara-module-developers/fd31e975-9f9f-47c6-b94a-e05757ad34b7%40googlegroups.com.

[Garrett Kinsman]

Guys,

What do you think of this Brillo Network? Google is supposedly working on and IoT network (whatever that means ;) Supposedly it will be an OS for small connected objects, but its potential for mesh could be very interesting. They're supposedly releasing it at IO. A few tweaks to the android OS, and how we think about the internet will change forever.... 

Looks like IO is going to be an interesting one. 

[Paul Gardner-Stephen]

Just a quick heads up on our work that might be relevant for this:  The Serval Mesh contains most of the software ingredients for creating these kinds of off-grid and strongly encrypted and authenticated networks.  There are some gaps, but it generally works.  If an ARA phone had, for example, an ISM band UHF radio, we could make very easy use of it to provide a variety of services.  We approximate this now with our Mesh Extender hardware that puts a UHF radio together with a battery powered WiFi access point, but are keen to make a single phone or phone-like device that contains it all.

Paul.

[Garrett Kinsman]

Paul,

Serval is very cool but I personally found the interface difficult to use. If only someone would create easy to use mesh software...

[Garrett Kinsman]

Paul, Did you mention you work on Serval?

Would be very interested in chatting. I've had some great conversations with Monkeybrins and Airjaldi. I think even a child should be able to easily build and deploy Internet infrastructure. 

I think ARA will make a great platform for Wireless internet and distributed computing... 

I was hoping they would allow replaceable screens,  which I would swap out for a 5.8 GHz sector antenna (much like the Ubiquity M5 antennas).

Let's chat