Wireless Mesh Networks for Internet of Things Latency
More bandwidth! Speedier connections! Greater accessibility! Bigger data faster! These are the demands of tomorrow's IoT. Can they be met? If so, how?
BOSTON – The issue of Internet of Things security has already been hammered to death on tech news sites (including here at Enterprise Networking Planet). Security, however, is far from the sole – or even necessarily primary – bugaboo facing widespread IoT deployment in the enterprise.
Accessibility – security's mortal foe – is potentially just as big a problem, according to pundits who spoke at this year's Connected Cloud Summit in Boston.
Bobby Patrick, chief of marketing for Hewlett-Packard's cloud division, discussed this very issue as he described the challenges of getting the healthcare sector on board with the Internet of Things in his keynote address at the Connected Cloud Summit.
"Wireless networks [are] insufficient] for [the] volume" of data healthcare providers handle, said Patrick. He added: "Sensors don't always work as advertised."
Laila Partridge, managing director of Code On Technologies and another speaker at the Summit, agreed. "Wi-Fi has pretty high needs for energy."
But even though healthcare and the life sciences deal with the biggest data out there, the problem is not limited to that sector only.
"[O]f course there ha[ve] to be some features...that don't exist in the consumer Internet," said Joe Salvo, manager of General Electric's Complex Systems Engineering Laboratory and founding director of the Industrial Internet Consortium, during a panel discussion at the Summit. "Guaranteed bandwidth, guaranteed latency."
Don Dodge, a developer advocate at Google, put the problem more succinctly: "How [do you] get the sensor...to the Internet when Wi-Fi is not available[?]"
Dodge then went on to answer his own question. "Well, you don't do it with cell, because 3G and 4G [are] just too expensive to send data[.] We need a very low-cost, low-bandwidth [solution]. Reliability doesn't matter; if I send my bit one minute from now or ten minutes from now or one day[,] it doesn't really matter… You really have to create an alternative network to do that."
Other presenters outlined their own additional requirements.
"One big challenge is availability," said Patrick. "You can't have maintenance…on Sunday night [in the healthcare context]. The infrastructure and the technology must survive failures[.] You need to be able to handle millions of nodes; you need…an x-compute big [data] algorithm."
And, of course, where big data and big data availability are implicated, so are big analytics. "Analytics and dashboards are required," specified Patrick. "You have to give other people access to it – cross-enterprise, cross-partner access."
Several solutions have been floated, most involving some form of a distributed mesh network. And a ubiquitous one at that.
"You need a mesh of computing both down on the streets and in your pockets," said Patrick. "We are putting this infrastructure, this distributed computing mesh – ...into...big companies around the world, and we're putting it into our customer centers."
The benefits of distributed mesh networks are manifold: they are cheap, they are easily set up, they are fast, and they can be private. That means improved bandwidth, improved latency, improved security, and fewer to no brouhahas over net neutrality.
Because of these advantages, activists have developed an enthusiasm for wireless mesh networks. Meshes have also proven popular in developing interconnectivity in more far-flung parts of the globe. Even worldwide wireless mesh networks have been envisioned as an alternative to the Internet proper.
"I'm very interested in alternative networks; I think it's gonna be a multi-billion-dollar opportunity," said Dodge. "Wi-Fi isnt the best solution all the time. There's an opportunity for an alternative network – even for indoor types of applications – that could be huge."
"One company I'm looking at uses these little home bases…to enable sensors in your home [and] also creates a mesh network of these home bases," Dodge told attendees. "You can use about 20 of these home bases to cover the entire city of San Francisco. The company can also use it to relay other messages from their sensors to the Internet. …With 10,000 home bases, you could cover 80 percent of the United States."
Cory Von Wallenstein, chief technologist of Dyn and member of the same executive panel as Dodge, reported that his own company was using this very technology. Other companies want to take this to the next level yet.
The major drawbacks of mesh networks, after all, come down to scalability and reliability: If one part of the mesh goes out for whatever reason, that leaves other parts of the network isolated from each other. That's where redundancy comes in for the mesh enterprise – lots and lots of redundancy. But how do you achieve that redundancy efficiently?
One solution for major carriers like Comcast being floated around is that of double-dipping in their customers' connections – a plan being charitably dubbed "Carrier Wi-Fi." The theory is based on the notion that a reported 70 percent of consumers have active Wi-Fi connections more than 70 percent of the time. As 2G (which many GSM-based machine-to-machine devices presently use) gets gradually phased out while 3G and 4G are proving to be highly expensive and highly burdened options for "always-on" IoT devices, developing a distributed mesh system by piggybacking on top of existing customers' Wi-Fi connections is becoming an attractive cost-saving measure.
Or, to put it another way: The peasantry's hands don't look full yet.
Dodge, as a Google representative, also favors this mentality of co-opting consumer devices – even mobile ones – and, with them, it would seem, consumer bandwidth and data plans. "What if you put an app on the smartphone that was simply a relay?" Dodge pondered aloud. "It can take information from sensors and relay it to the Internet."
Barb Darrow, senior writer for GigaOm who moderated the panel, responded to Dodge's vision by asking him the question that the members of the hoi polloi in the room were thinking: "Will you pay my phone bill?"
Googler that he is, the aptly named Dodge could only stammer, "Uh, the business model hasn't been worked out yet..."
The audience laughed – uncomfortably. Only time will tell what consumers will find themselves enmeshed in – and how they will react.
Photo courtesy of Shutterstock.
Joe Stanganelli is a writer, attorney, and communications consultant. He is also principal and founding attorney of Beacon Hill Law in Boston. Follow him on Twitter at @JoeStanganelli.