One of the advantages of Manhattan’s Union Square area is the high-concentration of coffee shops, startups, WiFi, and visiting San Franciso-based entrepreneurs. On a recent excursion into Manhattan, I took advantage of most of the aforementioned. I was meeting Micha Benoliel, a Co-Founder of Open Garden, at a Pain Quotidien–a French-inflected communal bakery and coffee chain that has invaded NYC–to discuss his mesh networking startup over espressos.
Open Garden was a crowd favorite at TechCrunch Disrupt NY, and though they didn’t win the final Battlefield jousting contest, they picked up a Most Innovative accolade.
While I was waiting for Benoliel to finish up a meeting, I was trying to log into PQ’s WiFi network, which I learned from the waitress was for employees only.
However, there was an open WiFi network at next door’s swanky ABC Furniture’s coffee bar–don’t ask, in Manhattan, furniture stores manage to dispense excellent coffee and tea. But unfortunately I didn’t have the password.
This is where Open Garden would help. Their idea is to transform the short-range Bluetooth transmitters on our smartphones and tablets into an ad hoc, multi-hop network. In other words, with enough Open Garden users in my vicinity at Pain Quotidien, my packets could jump onto another Garden member’s device, and then leapfrog from that device’s landing pad onto the public WiFi network.
It’s a fascinating (and complex) idea, in that the crowd becomes the network–it magically emerges. My interview with Benoliel was quite fun and geekily exciting–I hope this is captured below. Enjoy!
Your team has deep roots at BitTorrent, and I see some similar concepts in terms of sharing load across multiple device, but this time at the endpoint level.
“The are some similarities in terms of leveraging everyone’s device. I think what you’re getting at is a feature, which is not implemented but will be available in the coming weeks, for bundling channels. If you have multiple devices with direct access to the Internet within the mesh, then these devices will be available to bundle the channels in parallel.
A simple example of this is with a web page. One channel will be available to load the images, and another the CSS.”
I would have called this, based on my background in archaic telecom technology, inverse multiplexing. Doesn’t this approach require more infrastructure in the cloud?
“That’s a good word to describe this. The magic involved in this, though, is that it doesn’t require any equipment or software in the cloud. It does require every device in the mesh to have the application installed.”
So I’m getting a fatter pipe than I would have if I didn’t have Open Garden installed.
“Yes! We leverage the network–at the edge.”
Kind of amazing. Would the ISPs be able to figure out that these packets are from OG devices?
“From the network side, it’s very difficult to figure this out. The ISP would essentially not know.”
Let’s get back to the crowd. Your approach is based on multi-hop technology, which usually involves reducing the amount of radio interference What would be the optimal situation be for OG. Presumably, a crowded area with lots of people using the devices?
“First, the latest version of Bluetooth is low on energy consumption and it doesn’t drain your battery down. Second, we found a way to avoid discoverability and pairing. The app is almost independent from the wireless interface.
So yes, an event–stadium or just a conference–where very often the carrier network is getting congestion. Say among four or five carriers, maybe one would have a good network coverage and some of the users would be connected to a private WiFi network, which has capacity. If all the devices in this space had Open Garden, then Open Garden would figure out the best path to the Internet. If 3G or 4G is congested, it would push traffic to WiFi.”
So OG finds the path using just Bluetooth and the crowd doesn’t have to do anything other than being there.
“You are completely right. What’s implemented now is seamless discoverability, and what we call automatic path choice, which is to choose the best offramp to the internet, and mutli-hop.
Multi-hop is very interesting to carriers, because right here in Pain Quotidien, there might be WiFi, but only a few people are dedicated to that WiFi. If they have Open Garden, then I could hop from the phone tethered to my laptop–if I couldn’t get a signal–and connect to WiFi.”
Of course, this is just what I could have used today, since Pain Quotidian’s WiFi network is private, but the coffee bar next door has an open network. So this all depends on enough people having Open Garden on their devices. And this leads to kind of an obvious statement in the form of a question: You’re very dependent on scaling effects?
“We are and we are not. We are not because you can already benefit from Open Garden just by using it to share your mobile data plan among your devices–that puts $20 to $30 back in your pocket. Obviously, we need more people to start benefiting from network effects. On the other hand, you just need one person within Bluetooth range to benefit from Open Garden.”
We could have all used OG at TechCrunch Disrupt since one of the WiFi bands, as I recall, was completely congested and few could get access.
“Yes, Open Garden would have figured out where the network is not congested and connect you to that part.”
Is OG dynamically slicing up the available bandwidth?
“We tend to prioritize your traffic. If you are using Open Garden but you are the main offramp to the Internet, then your traffic will be prioritized so you don’t suffer from others piggybacking on your device.”
Is the mesh completely anonymous or am I aware of others?
“You are part of the mesh, but you will see the names of the devices.” [see pics above]
I was a little curious about the workings of the algorithm–about how it decides when to abandon say, an intermittent WiFi connection and switch to OG.
“The algorithm will try to optimize your capability, but not only for speed, but for less expensive routes.”
Can you reveal more about your roadmap for features?
“We started with a basic notion of being completely open. We had to go public because to improve the software we need a lot of users. We were running in private beta with a few hundred users but that was not enough to improve the software. Since the launch at TechCrunch Disrupt, we had 50,000 installs on MacBook, Android, and Windows together. It’s a good number.
When we reach critical mass, people will say ‘I don’t want someone piggybacking on my data plan. I want to decide who can access my data plan.’
This will require a lot of iterations. We will probably enable you to connect through Facebook Connect or Google Plus and then let you decide whether you want to share with everyone, friends of friends, just friends, or only you and your devices, and how much data you want to allocate to Open Garden.”
Have you thought about an Open Garden economy?
“We thought about an ecosystem where to incentivize people to share with others they would be rewarded with credits. This will require iterations before we arrive at something that works perfectly.”
Have you had any talks with the carriers or other ISPs?
“We have an agreement with one European carrier. The first reaction is usually we don’t want you to exist because you are stealing our bandwidth. We don’t because we’re not selling it! It’s a normal reaction because this is really a disruption.
For carriers who have a lot of experience with mesh or at least did a lot of research–like the one we signed the agreement with in Europe– really understand the benefits: enabling WiFi offload in more places, increasing the range of WiFi through multi-hop, and also improved performance for video apps, which require lot of bandwidth. Being able to bundle channels for video is interesting to the carriers. On a single device we can bundle WiFi and 4G for greater bandwidth share.
In a way, this should be seamless for the user, you should be able to connect if there is a network available, regardless of whether it’s AT&T or Verizon, WiFi or cellular. To some extent, we could also decide to become an MVO–in a few years. We could sell to the user data that is available in time and space. Our plans are open.”
Finally, do you have a number of subscribers you’d like to reach in say a year?
“Hopefully, within two years will reach two to three million users. And if all go goes well, maybe there’ll be 10 million!”
Micha, thanks for your time. Speaking for lots of people, we’re all looking forward to being part of an Open Garden mesh soon.