 |
Robotbase 22 January 2015 - by Maxine Bingham, Editor-in-Chief
|
You’re
asleep in bed, when Rosie the Robot slides over to you to wake you, turn on the
lights and ask if you want your coffee started. She then reminds you of the day’s
appointments, as well as the latest football scores. Thanks to Rosie, the
coffee is made, the right lights are turned on and the temperature is adjusted
to your liking by the time your feet hit the floor. She looks at you and asks
if you’d like some jazzy music as you look tired. You tell her “yes.” While you’re away,
she turns down the thermostat and adjusts the lights. At night, when you are
exhausted, she reads a story to your child before bedtime.

This is not just a scenario out of the Jetsons’ cartoon series, it is a
new implementation of the Internet of Things integrated with artificial
intelligence, speech recognition, natural language processing, emotion, object
detection and facial recognition and neural networks, to bring to our homes a
Personal Assistant that can anticipate our needs and handle connected devices
in one place. She (or he – you can choose your avatar) can even show you
recipes as you cook and ensure the door is locked and oven turned off while you’re away
automatically.
CEO
Duy Huynh, originally from Vietnam, founded New York-based Robotbase, the
company that has designed this Personal Assistant robot. Robotbase announced it
at International CES in early January.
Huynh
designed this invention after analyzing problems he found with the Internet of
Things. He saw two main problems. One,
IoT solutions amass a great deal of data, but companies don’t know
what to do with it in order to create better customer experiences, and, two,
people today need separate apps for each connected device, which can quickly
get out of hand.
The
video on the company’s KickStarter page provides a number of
scenarios, including taking photographs on the robot’s own initiative. One can
see the great interest in this invention, as they have
greatly
exceeded their goal of $50,000. As of January 20, they’ve raised $119,825. The
campaign will end on February 5, 2015 at 3:39am PST.
Robotbase
did collaborate with Samsung on a software component that is the foundation of
the robot software today, but won’t say if Samsung is an investor. A small team
(under 15) of software engineers, hardware engineers, mechanical engineers, and
product designers have created the Personal Assistant (to make it feel more
personal, owners give her/him a name).
The
Internet of Things, or the IoT, is a catchall term for devices that are
connected over the Internet, designed to be smart and to use as little human
intervention as possible. In the case of Robotbase, speech recognition is
performed by an internet-connected server whereas the natural language processing
is done locally. Other IoT examples include the Nest thermostat and the Fitbit
fitness tracking “wearable.”
The
IoT is considered the greatest revolution in technology since the Internet. The
possibilities for positively affecting our lives are endless – from connected
cars, including those that drive themselves and smart cities that reduce
congestion and pollution, to eHealth applications that allow doctors to
diagnose remotely and energy management systems that reduce energy consumption.
Robotbase
plans to offer its Personal Assistant in Q1 2015 at a retail price of
$1,995.00. |
Photo of Robotbase CEO Duy Huynh and the Personal Assistant courtesy of Robotbase |
|
 |
Roost 28 October 2014 - by Maxine Bingham, Editor-in-Chief
|
Startup, Roost, announced their first product on 21 October 2014, a retrofit "smart battery" for the home. Coupled with an
intuitive smartphone app, the Roost Smart Battery delivers
notifications to your smartphone when the smoke alarm activates, even
if you are not at home.We spoke with CEO and co-founder, Roel Peters, about his new venture.
"Simply download our Roost app, connect to your home WiFi, and then
insert the Roost Smart Battery in an existing smoke alarm to appreciate
how simple and intuitive it is to install and use," Peters said. "No hubs, no wires, and no pricey new
devices. Our retrofit approach to the smart home shows that peace of
mind doesn't have to be expensive or complicated. We've chosen to apply
the Roost Smart Battery to smoke alarms first because quite simply they
are the most important devices to ensure the safety of your home and
loved ones."
The
patent-pending Roost Smart Battery combines a lithium
battery and advanced Wi-Fi in a familiar 9V package. According to the company, it will begin
shipping in Q2 of 2015. The company plans to offer pre-orders through a
Kickstarter campaign in the next few weeks. They also claim that the battery will, although dealing with increased performance, last five years.
Sounds intriguing. We wish them luck.
|
Photo of Roel Peters and Roost SmartBattery Alert Courtesy of Roost |
|
 |
Seed Labs 14 December 2014 - By Ron Bingham, Technology Editor & Sr. Analyst
|
Startup Seed Labs, whom we met at DEMO Fall 2014, provides a
design kit for manufacturers to easily incorporate IoT capability into everyday
products.
Initially, the company’s design efforts were directed at
providing software to manage and coordinate existing home IoT products. After
examining the number of different home IoT products available, they concluded
that there were too few products available to make their original control hub
idea viable. Seed then decided that what the world needed were more devices in
the home that are IoT-enabled, thus they pivoted their design strategy to make
it simple and easy for device manufacturers to incorporate IoT capability in consumer
products.
Seed’s prototyping kit offering includes
three Seed microboards, as shown below, and smartphone apps for iOS and Android
that communicate with the Seed BLE microboards to test and verify design
validity.
Features of the design kit include:
- Up to 1,000 ft. range (line of sight).
- Ability to group and control a set of
devices.
- Multiple smartphones can access and
control all devices.
- Public and private modes, i.e.,
Seed-enabled devices can be set to restrict access to specific smartphones or
allow access to any smartphone within range.
- Smart phone apps automatically detect
Seed enabled devices’ characteristics and build on-the-fly
appropriate controls.
- Automatic detection pairing of
Bluetooth-device-to-smartphone.
- Automatic user identification based on
smartphone or fob carried by user.
The innovations
How Seed Achieves Extended Range
As can be seen on the microboard, the larger Bluetooth Low Energy
(BLE) SoC (system on a chip) is combined with a smaller amplifier chip to
increase receive sensitivity and extend broadcast range. Also, notice the
antenna trace on the periphery of the microboard. These design choices combine
to extend range and increase receiver sensitivity to 108 dB.

How Multiple Smart Phone Access is Achieved
Seed uses a modified Smart Bluetooth protocol that turns
Bluetooth from a star network to a mesh network. So, instead of a single
smartphone connecting to multiple devices, multiple smartphones can connect to
multiple devices. Mesh networks also extend the physical range limitation by
allowing a smartphone to connect to the nearest device to be able to access all
of the devices in the mesh network no matter their location.
Support for IoT Device Manufacturers
For a device to be convertible it must have an embedded
microprocessor that can talk to the Seed IoT micro-controller. Most appliances
today already have microprocessors performing control functions. Failing that,
the simplest alternative implementation is for the manufacturer to supply a
Seed-enabled wall power plug that can turn the appliance off and on from the
Seed smartphone dashboard.
In addition to the design kit, Seed provides technical consulting
and support from design through manufacturing.
Seed Labs
was founded in 2013. They raised $1M in angel funding in September 2014. The
company includes 50 software developers, hardware engineers and UX designers.
They are based in San Francisco (sales, marketing and support), Krakow, Poland
(where they started and have R&D and engineering) and Guangzhou China (Asia
engineering and sales).
We find Seed’s technology and B2B strategy very
compelling. We’ll be watching for their news.
|
Images Courtesy of Seed Labs |
|
 |
SIGFOX 7 July 2015 - by Maxine Bingham, Editor-in-Chief
|
SIGFOX provides a cellular network
on top of the Internet that delivers a complementary forth connectivity
solution that offers IoT connectivity at a lower price and much longer power battery
life than WiFi or Bluetooth, for example.
They achieve that with a
solution that allows you to exchange small packets (messages) via their own
antennas/cell towers over unlicensed spectrum (via partners). Thus they can
cover wide areas and sell connectivity at a very low cost. SIGFOX’s
connectivity can enable devices to have 20 years of battery life, for a price
of about one dollar per device per year.
According to Marketing Chief
Thomas Nicholls, SIGFOX democratizes connectivity.
The company began in France,
since the founders and management team were there. Today, SIGFOX has 450,000 square
miles covered in the country. They have 8M device connections and several
hundred thousand devices on its network.
The company raised $115M to
cover all of the US and finish Europe, and to start covering Asia and Latin
America in February of 2015. According to TechCrunch, this investment was by a
number of strategic investors including the Spanish, Korean and Japanese
carriers Telefonica, SK Telecom and NTT DOCOMO (whose Ventures arm did the
investing), and French industrial giants GDF SUEZ, Air Liquide and
Eutelsat. Also participating was Elliott Management Corporation.
Previously, SIGFOX had raised about €30 million from investors that include Intel
Capital, Partech and several other French investors. On 15 June, the company announced
additional investment (amount undisclosed) by Samsung Investments.
Said Thomas, “The goal of
SIGFOX is to finally enable what we believe what the IoT is, since, today, there
is no way we will get billions of devices connected. SIGFOX removes the barrier
to entry for the IoT to happen.”
In terms of monetization,
Thomas said that, “We sell subscriptions to partners, we’re like an AT&T. We
sell data subscriptions that provide access to the pipe, we don’t build the
solutions. In parallel, there’s an ecosystem of SIGFOX-ready device, thus we
spend a lot of time working with device manufacturers to create a large catalog
of SIGFOX-ready solutions that others provide to their customers.”
The company plans to roll
out a new version of its web site with a fully-fledged catalog of SIGFOX-ready
devices. Currently there have 60-70 off the shelf compatible devices. Some
partners provide SIGFOX-ready devices directly to their customers, so that
those would not appear on the catalog.
SIGFOX offers a free a
software protocol and a SIGFOX-ready certification program.
Thomas noted that, “We work
with manufacturers to ensure optimal performance.”
Device manufacturers pay SIGFOX a yearly subscription; it is up to the partners how to charge, or if
not to charge, for their SIGFOX connectivity. Some may bundle in the cost via
the lifetime of the product, for example. Thus, the end user interacts with the
device manufacturer, not with SIGFOX directly. “It’s a B2C solution,” added
Thomas. For example, a farmer might monitor water in the soil and via a company
providing such a solution to the customer, so the end customer focuses on the
solution and not how SIGFOX works.
According to Thomas, “SIGFOX
was born by a team of people who had been working in connectivity for a few
years and who came to the conclusion that there was a problem with IoT
connectivity. Connectivity wasn’t good
enough for all those use cases where you just need to extend the five senses of
the human by collecting a small amount of data about the physical world around us.
What’s unique in our approach is rolling out a new operated cellular network,
no one thought of doing that, and for established communications companies it
would have meant they needed to buy expenses licenses to operate, and thus
billion of dollars of investment.”
SIGFOX signs an exclusive
partnership deal with each partner in each country, it’s like a franchise
Thomas told us, and where that company becomes the local operator. The logic in
SIGFOX’s network is running in SIGFOX’s cloud and their bay stations. All
devices run through the SIGFOX cloud, that makes one network. According to
Thomas, “We break down the barriers of roaming. You can connect from anywhere.”
If a major entity, such as a
city wants a SIGFOX network, they can deal with the company directly. Covering
a city depends on the size of the city. “It wouldn’t require that many cells,
it’s very quick and not that expensive to cover a new area,” explained Thomas.
Bottom line, SIGFOX is
designed for use cases where there are small messages being transmitted. It is
a clever and increasingly successful approach that is aimed at connected those
million, or billion, of devices cheaply and with very long battery life.
|
Photo of Thomas Nicholls from the Internet
© 2015 IoT Perspectives |
|
 |
Strawberry Energy 18 November 2014 - by Maxine Bingham, Editor-in-Chief
|
Is there a smart urban “tree” in your future? I had the good
fortune of speaking with Miloš Milisavljevi?, founder and chief executive of
Strawberry Energy Company during the RE.WORK IoT Summit in San Francisco. He is
the inventor of the “Strawberry Tree,” a solar-powered “smart city” platform
for urban environments.
The black Strawberry Trees are comprised of a wooden bench
and a beautiful steel sculpture in the shape of an abstract tree with a
“canopy” of solar panels to provide free Wi-Fi, device charging, environmental
sensing and, a place for community gathering.
The Strawberry Tree’s patent-pending technology includes:
- 530W solar panels
- 4.7kWh rechargeable battery
- 16 (special to the tree) charging cords – so don’t take
them, they won’t work!
- Up
to 1.5M 10-minute annual battery charges
- 6.2k 10-minute charges without sunshine
The solar panels do not move with the Sun, but are optimally
placed according to engineering calculations, toward the south, with the angle
depending on location.
Some 350,000 people are using Strawberry Trees, at first in
Eastern Europe, which makes sense since the company is headquartered in Belgrade,
Serbia, but is planning to open an office in the U.S.

The company’s articulate CEO, Miloš Milisavljevi?, told us, "I wanted to put
renewable energy sources into people's everyday lives, and realized that the
best way to show people benefits of these clean energy is to solve their
problems with the usage of green technologies in public spaces. On the other
hand, with over six billion portable devices around the globe, energy and
Internet have become 21st century water and air. Still, people get disconnected
every single day. I realized that the time has come for our public spaces to
change and adapt to the mobile generation of our time.
Miloš
added, “That’s why we created our Strawberry Tree, conveniently placed all
around the city to answer the modern “thirst” for power and connectivity. Just
like every tree in the nature, that grows and has new branches, our Strawberry
Tree grew into a smart city platform to offer new services to people. Our
future plans are to further expand into the US market, because we want to
enhance the smartness of public spaces across America."
Strawberry Trees are in 12 locations so far, in nine
different cities in two European countries, and very soon the thirteenth will
be installed in Timisoara, Romania. Exact locations are mainly in parks and
city squares, to date. The company is also pursuing private lands such as universities,
business and other campuses, as well as spaces in front of shopping malls,
etc. The company has signed a Letter of Interest from California Community
Colleges’ Chancellor's Office, which represents 112 campuses, which I think is
quote impressive for an early foray into the US market.
Privately-held Strawberry Energy was founded in 2011. So
far, they have received 100K Euros (approximately $125,000 USD) funding by Bulgarian Accelerator Venture Fund Eleven.
According
to the venture fund, “Our
Strawberry Energy startup is the first startup from Serbia to receive 100K
investment from Bulgarian Accelerator Venture Fund Eleven in April this year.
From a total of 276 applications from 26 countries, 11 companies are chosen for
the investment, while only five startups received 100K and six startups
received 25K.” The company is in the process of raising additional
investment, especially to help with expansion into the U.S. and other geos.
I am impressed by the company’s execution and track record,
as well as its founder/CEO and its vision for making urban environments more
enjoyable, greener and practical for a connected age. I can see these in dog
parks, high schools, medical facility and research institution grounds, and
other open spaces. And, hopefully graffiti, vandalism and theft of cables,
etc., can be addressed to keep Strawberry Trees in action.
|
Photos Courtesy of Strawberry Energy |
|
 |
Sunrise Micro Devices 19 February 2015 - by Ron Bingham, Technology Editor & Sr. Analyst
UPDATE: On 16 April 2015: ARM announced the acquisition of Wicentric, a Bluetooth Smart
stack and profile provider, and Sunrise Micro Devices (SMD), a provider
of sub-one volt Bluetooth radio intellectual property (IP). Terms of the
agreements were not disclosed. The IP of both companies will be
integrated to form the ARM® Cordio portfolio.
|
What do you get when you mix one
part of the world’s leading supplier of semiconductor core IP with three of the
leading radio scientists and engineers from Motorola Labs? It’s called Sunrise
Micro Devices, Inc. (SMD) of Deerfield Beach, FL. Using a proven business
model and an undisclosed amount of investment from ARM Holdings, SMD is the
first venture to offer a complete Bluetooth Smart core IP radio with baseband,
the CORDIO BT4.
Using SMD IP, Internet of Things (IoT)
designers can easily and quickly include the latest Bluetooth capability in
their chip designs. The radio consumes less than 6.5 mW when transmitting and
only 700 nW in sleep mode and costs less than if designers did their own radio
from scratch (no easy task).
According to Prithi Ramakrishnan,
SMD Product Manager, Marketing and Strategy, “A company could spend tens of
millions to develop a Bluetooth radio design whereas we offer a superior,
state-of-the-art, FCC and Bluetooth-certified design for a onetime licensing
fee and a per part royalty.” Ramakrishnan declined to give specific pricing but
did say, “We believe the market demands a price of under a couple of dollars
for IoT Bluetooth Smart functionality.”
The tech-savvy founders and the troika leading SMD are (right to left below) CEO Fred
Martin, PhD EE, CTO Ed Callaway, PhD in Computer Engineering, and Chief
Scientist Paul Gorday, MSEE, who have worked together on radio science and
technology for several decades at Motorola Labs. Between them, they hold 107 issued
patents, authored scores of technical papers and books, and have participated
in the development of the IEEE-802.15.4 radio-networking standard. Callaway is
also an IEEE Fellow and a member of Bluetooth SIG’s
Core Specification Working Group.
 
SMD
started as a consulting company in 2009, but migrated to a product company in
2014. The company is under 30 employees.
The
driving force behind their business idea is that the Internet of Things is
requiring low-cost, low power combinations of microprocessors, sensors and
interconnectivity. One way to achieve that goal is to design chips that combine
all three. SMD is supplying the interconnectivity portion with its Bluetooth
Smart IP core.
Features
The CORDIO BT4 has some interesting and unique features (reference design below). It can
operate from a sub-volt power supply and accepts voltages up to 3 volts for
increased transmit power. It has extensive power management features including
a 700 nW sleep mode. This collection of features is ideal for the IoT
environment, particularly for applications that require battery power or energy
harvesting. Applications such as beacons and distributed sensors fall into this
category. The CORDIO BT4 can operate up to 10 months on a 70¢ hearing aid
battery less than 12mm in diameter.

The digital radio receiver
design uses a single phase-locked-loop and mixer to generate an intermediate
frequency that is low enough to be converted to a digital signal. From here,
all of the processing is done in the digital domain including baseband, link
layer and a digital host processor interface that makes software development
straightforward and efficient. The receiver’s sensitivity is -94 dBm which means
it can detect a signal that is a half of a trillionth of a watt. The transmitter is an FM modulated PLL followed by the PA.
We
think that Sunrise Micro Devices is a player to watch once they have mastered
the Bluetooth radio arena. In addition, the company’s basic design can be
relatively easily extended to other radio based technologies such as WiFi. With
a proven ARM-style business model, ARM investment and a stellar technical team,
they could be poised for dominance in this important IoT technology market. |
Photos of Prithi Ramakrishnan, Fred Martin, Ed Callaway, Paul Corday and CORDIO BT4 Reference Design Courtesy of SMD |
|
 |
Thalmic Labs 6 October 2014 - by Maxine Bingham, Editor-in-Chief
|
Editor-in-Chief Maxine Bingham interviewed ThalmicLabs' Evangelist Chris Goodine at the Designers of Things conference in San Francisco in September. ThalmicLabs closed a $14.5 million Series A funding round in 2013, led by
Spark Capital and Intel Capital, as well as individual investors such Marc Benioff, etc. Below is an edited Q&A. BTW - the definition of thalamus is "the middle part of the diencephalon (posterior section of the forebrain) through which sensory impulses pass to reach the cerebral cortex."
MB: Tell me about your company and product.
CG: We're hoping to expand Myo (note: gesture-based armband) into a catalyst for a bigger goal, which is to improve human-computer interaction in a world of wearable and ubiquitous computing.
We wanted to combine mobile and gesture control in our Myo armband.
When we launched the video on the web site, we showed what the world
will look like with this kind of capability. We're based in Canada and
we're planning to manufacture in North America.
MB: Is Myo shipping?
CG:
Developers and end-users can pre-order it now for $149. In fact, 40,000
have already been pre-ordered. Developer kits are available now. We're
also building some applications ourselves. We expect to have a consumer
launch within a month or so. We're focused on the consumer, to have Myo work right out of the box.
MB: What kind of applications are possible with Myo?
CG:
Some examples include using gestures to underline, zoom in or create
circles when giving a PowerPoint presentation, you could pause or
fast-forward a film on Netflix sitting in your chair, there's also the
ability to control music on the go when you're skiing, snowboarding or
running, as you don't have to take a phone out of your pocket, you use
your hands.

MB: How does that work?
CG: The
core of Myo is muscle sensing. Based on your muscle activity we
understand how you position your hands. No one else is using electrical
muscle activity like this. Our solution also uses a unique pinky tap to
remove false positives that would otherwise result when hands are used
for everyday actions, such as writing, waving, and holding things. The
technology used to measure electrical activity on the skin has been
around for years, but current sensors have picky environmental
requirements, so we built the sensor ourselves. Our patented
"secret sauce" is that we can see the signal among the noise. We process
on the device, so it's faster and has an efficient battery life.
MB: What about compatibility?
CG:
Myo works with Windows, Mac OS, iOS and Android. There is an SDK
(software development kit) and APIs (application programming interface)
for developers to create on platform(s) of their choice. It's publicly
available, you can just download it.
MB: How many degrees of freedom does Myo offer - beyond up and down?
CG: We
offer nine via a 3-axis accelerometer; 3-axis gyroscope and 3-axis
magnetometer (compass). We use sensor fusion to combine the outputs of
these sensors to give orientation in the form of roll, pitch and yaw.
MB: Why did you choose a gesture-based solution?
CG: It's the most natural. Mobile touch is limiting and voice can be socially awkward.
MB: Thanks, Chris, for speaking with us.
CG: I enjoyed our talk very much.
|
Images of Chris Goodine and Myo Courtesy of ThalmicLabs |
|
 |
ThetaRay 8 May 2015 - by Ron Bingham, Technology Editor
|
It’s one
thing to detect and combat a known attack, but, what do you do when your
company is facing the unknown? We had a conversation with Mark Gazit the CEO of
ThetaRay, an Israeli company that claims to have solved the difficult problem
of detecting previously unknown threats, and enabling enterprises to quickly
combat them.
IoTP:
Tell us about ThetaRay.
MG: We have one of the
largest security companies in Israel. We are venture-backed. [According to the Wall Street Journal, the company raised $10 million
in 2014. Participants in the round included existing investors in the company—General
Electric Company, Jerusalem Venture Partners and Poalim Capital Markets Ltd.—as
well as new financial and strategic investors. GE is a customer, as is Bank Hapoalim, Israel’s second largest bank.
Founded in 2012, ThetaRay spun off from the mathematical research of the
company’s founders—professor Amir Averbuch of Tel-Aviv University and Yale’s
Ronald Coifman. The product, deployed on site or run off a cloud, is based on advanced
mathematical algorithms that learn the “normal” behavior of a system, and flag
anomalies, like a cyber-intrusion, when they occur. ThetaRay started out as one
of the companies in Jerusalem Venture Partners startup incubator in Jerusalem.
“Our information security product helps enterprises identify unknown threats by
analyzing different streams of data from all around the organization,” ThetaRay’s
CEO Mark Gazit said.]
We have offices in
Jerusalem and Tel Aviv and are in the process of opening an office in the US.
IoTP:
What’s the impact of IoT on security?
MG: The world has really changed. In 20 years,
there will be 4.5B Internet users and 25B IoT devices. If the iPhone is a
device – then [that figure] is an understatement. Suddenly it’s 1 trillion
instances today talking to each other. Firewalls, access lists and anti-viruses
don’t work in this new world of the IoT. The only way to protect is for
machines to protect humans.
IoTP:
So, in effect, every app has to have its own immune system?
MG: It’s
amazing what you said. Yes. 20 years ago people
believed if the environment was clean and you washed your hands used
antibiotics you’d be safe. Now there are virus-resistant antibiotics, and it’s
harder to keep a clean environment. So, we need to strengthen a device’s immune
system. Now, for example, you can take any managed refrigerator and make it
into a weapon.
IoTP:
How do you prevent that?
MG:
Through algorithms. You need to analyze data in real time and note unusual
patterns. You have to detect and then can protect. You need to know if you need
to shut things down, or, if it’s only an alert. Making a “clean” environment
just doesn’t work anymore. Every organization, especially when deploying IoT
needs a detection system.
However,
think of machines, they’re full of opportunities for malicious attacks. Human
beings can’t cope with it any more – that’s why we take a mathematical approach
that listens to the data.
When
something not normal is happening our system alerts the customer and helps them
understands what’s the problem. Thus,
our system is better equipped for large organizations and large amounts of
data.
Our
approach is similar to the new world of trading. Before you had to be a good
trader, now it’s automated, people let machines trade on their behalf. Call it
the “trading of things.”
We apply
physics – diffusion
technologies, diffusion equations.
IoTP:
What data do you look it?
MG: That’s
the cool stuff. We protect financial institutions. Think of an ATM as a thing.
It’s connected with the Internet. You can now manipulate it to withdraw money
without even touching it – that’s a real thing happening now. I can make a
computer open an ATM door, you just have to move another motor to spin the
wheel to take the money. So, we look at the financial and hardware transaction,
first building a picture of normal activity. If money is taken out without a
card, we identify that.
IoTP: So
you collect every transaction?
MG: Yes,
every transaction, every piece of data.
IoTP:
You have to develop this hyper-dimensional view when I’m not being attacked?
MG:
Exactly. That’s why we work with relatively clean environments – if it’s more
than 20% dirty it’s too noisy an environment. Thus we protect enterprises like
banks and power plants.
IoTP:
What are organizations’ concerns about deploying security for the IoT?
MG: One
of the reasons for this resistance to security is the fear that false positives
create. With Target, for example, executives were fired because although a
security vendor discovered the problem they were ignored. Every day Target had
1500 alerts – low security alerts. But, who can look at this level of alerts?
ThetaRay
makes a better system because it uses math and physics – there are very low
false positives.
IoTP:
What’s your attack discovery process?
MG: By
developing parameters. Give us all the parameters and we’ll detect what’s
important and what’s not.
IoTP:
What’s better about your discovery process?
MG: It’s
unique in the industry. We have patents on all this. Our customer and investor
GE had checked the market, and they said our solution is unique.
It's exciting to contemplate the possibility of determing the "unknown unknowns" by exquisitely defining what is "known normal."
|
Photo of Mark Gazit from the Internet ©2015 IoT Perspectives |
|
 |
TreasureData 10 October 2014 - by Maxine Bingham, Editor-in-Chief
|
We had a substantive conversation with TreasureData Director [Note: subsequent to this interview, Hannah was promoted to VP] Hannah Smalltree about the company’s “managed
cloud service” (similar to SaaS) for data generated by the Internet of Things. The company came to our attention based on their blue chip investors, including Tim Guleri, Sierra Ventures; Yukihiro
Matsumoto (“Mats”), inventor of Ruby (my husband’s favorite programming
language); Dan Scheinman, Former SVP, Cisco, and
on the board of Arista Networks; Jerry Yang of Yahoo! fame, and others, as well
as the fact that they have so many case studies and customers. Cloud for the IoT can be hard to understand, I think Hannah did a great job of helping to explain it and what they do specifically. Below
is an edited Q&A.
MB: Some IoT industry
watchers say that because data processing will increasingly be on the
input/output device, especially in the consumer IoT space, that cloud computing
for the IoT is less important. HS: The data coming from IOT
devices is so massive, yet so important, that cloud computing is most
definitely an important component of the Internet of Things. Ultimately data
processing will happen in two places. Data needs to be collected, aggregated
and analyzed across all devices to learn how devices are performing. Companies
look for common usage patterns and common errors to determine what kinds of
data processing or automation should happen on the device. Then, there will also
be automation that happens within a device itself, for example, “when this
error condition occurs, take this action.” But to determine what that
automation will be – i.e, which error conditions are important - there still
has to be that overall analysis of data from all devices and that’s an ongoing,
iterative process. That data analysis of all devices is the piece that people
are interested in doing in the cloud for efficiency reasons.
MB: Explain what
TreasureData has been developed to do. HS: We’re in the Big Data space. We refer to the “big
data” that we collect, manage and allow customers to quickly and easily analyze,
as event data or “actions” – that is logs of how a user is using a device
versus more traditional structured customer records. We can collect thousands or
millions of these “events” from a device or server, enabling customers to get
value out of all that activity.
The goal with TreasureData is that customers
don’t need to worry about database management or configuration. We enable
customers to focus on their business, not the complicated problem of
integration between data (or “events”) collection, aggregation and management. We
can return data in near-real-time, so customers can analyze data from one to five
minutes ago. Customers need this kind of information to improve products, enhance
customer satisfaction, make their operations more efficient and enable new
business models – it’s disruptive in a good way.
We’re a managed service, so a customer doesn’t need to be an experienced
cloud or big data expert to use it. We provide complete end-to-end technology
capabilities, platform management and support for one monthly subscription
rate, similar to software-as-a-service. The idea is that with TreasureData, a
company can focus on analyzing data without worrying about database
administration or the underlying infrastructure.
MB: How does TreasureData
work? Our Treasure Agent (small software app) collects
data from servers or devices and streams it in near-real-time to the cloud. In
the cloud, data is stored in our Treasure Plazma system and our staff manages,
monitors, scales and secures the data. Then customers access our analytics
engines via a simple web browser. There, they use our Treasure Query
Accelerator to run queries and analyze their own data with SQL or customers can
also use their business intelligence/analytics tools to access data stored in
Treasure Data. Customers can store all their data in Treasure Data or we can export
query results or datasets any time. We leverage Amazon as our platform in the
US, but we can run on a variety of cloud platforms.
The Treasure Data
Service is based on a differentiated technology stack and unique database that
we've developed: Plazma is a distributed, object-based database that is purpose-built
to analyze massive data volumes. It includes a columnar layer for fast
analytics and archived data and a row-based layer for real-time data, plus the
required merge logic. The entire service is designed for exponential growth,
with a loosely-coupled, multi-layer architecture that allows I/O, processing,
and storage layers to independently scale and gracefully handle the
ever-increasing volume and variety of data our customers upload to our system.
MB: How much Big Data are we
talking about? HS: Today we manage over 9.5
trillion records (what we call “rows”) and that’s growing every second. We also
offer a free service so that customers can give us a try and can manage 1B rows
a month. In July 2013, we had over one trillion rows in our service. By November
2013, we had two trillion. Now, we're approaching 10 trillion. We're adding about
five hundred million records per hour, and that number is growing daily. Unlike
many MPP databases and Hadoop implementations, our architecture separates
storage from computing. Hence, based on our customers' collective needs, we can
scale up/down our storage and computing capacities independently from each
other.
MB: What kinds of value are
we talking about? HS: For example, customers can
learn a lot about how their products are behaving in the real-world. They can
see how they are technically performing and understand how customers are using
the products. This helps them design better, more competitive products.
Companies can also do things like preventative maintenance, where they can set
up rules for early intervention. For example, analysis might show that there are
often five common events that occur in succession before X becomes a problem,
so they can set up intervention after three of these events occur and prevent
the problem. However, before customers can realize the benefits from such
business alerts, they need a secure and fast way to aggregate all the data
being transmitted, and a platform that integrates data collection with their
analytics.
MB: Do you have any customer
case studies you can share? HS: Yes, Pioneer, for example, which many
of us know as a stereo company, is also invested in automotive telematics. They
are the number one auto electronics provider. They have a new technology and
service that collects sensor data from cars and sells that data back to the car
manufacturer, in fact, we helped co-develop the data collection technology
that goes on the electronics boards in the cars. This enables car manufacturers
to add value for their customers, with new services such as maintenance
reminders that are linked to your calendar and send an email to make the
appointment. The car can also alert you to potential problems or failures
before they appear such as “you need to get your brakes replaced.” Insurers could
potentially also use this auto data to help customers reduce their premiums, like
the Progressive insurance commercials talk about.
MobFox is one of Europe’s leading mobile advertising networks, with
offices in the U.K. and France. It’s a one-stop mobile advertising platform
that allows advertisers to run intelligent, optimized campaigns across a
network of 20,000 mobile publishers and many of the world’s largest third party
exchanges. In two weeks they were up and running and now have a complete data analytics
infrastructure that handles everything from data collection to storage and
analytics, that can scale to handle billions and billions of records being
added to a system in just one month.
There
are a lot more case
studies
on our web site.
MB:
What about security and privacy of all that information? HS:
While we manage and constantly monitor the service to ensure maximum security
on our end, it’s up to the vendor to negotiate their security and privacy
policies with their users. This is true for all IoT providers. It’s a two-way
street.
MB: Tell me more about the
company. HS: We launched our service
in 2012 and already have over 100 corporate customers and thousands of free ones. As you know, we’re backed by top VCs and angel
investors. We’re based in Mountain View, CA.
MB: Thanks for speaking with
us, Hannah. HS: My pleasure.
|
Hannah Smalltree Photo Courtesy of Treasure Data |
|
 |
TruSource Labs 20 April 2015 - by Maxine Bingham, Editor-in-Chief |
I began my career in tech as a Support Specialist for Fortune
500 companies in the US, Canada and the UK for what became D&B Software.
Thus, support has continued to be top of mind, and especially now as I wrestle
with IoT devices and need some help! So, was delighted to connect with a
company based in Austin, Texas, TruSource
Labs, that is focused on providing outsourced end user support for IoT
companies, starting with their first customer, Nest.
According to straight-talking co-founder, Alton Martin, privately-held
TruSource saw an opportunity two years ago for providing advanced end user support
functions for IoT companies’ customers. In their conversations with Nest, they
saw a raison d'être for a call center business that would be able to provide
sophisticated technical support, not just for Nest’s customers, but also for
IoT companies as a whole. This has led to fast growth. They grew from 0 to 325
people after 18 months and will soon add a second call center in Austin. They
plan to go global with call centers in Dublin, Ireland, to handle Western
Europe, and next, Singapore, for Southeast Asia support.
In addition to Nest, TruSource is handling outbound end user
support for a leading cloud storage company (you’d know the name), as well as
smaller IoT companies including several lighting control ones. They are also working with a
firm out of Melbourne, Australia, that provides geo-locating IoT so retail
companies can offer discount coupons to shoppers as they shop. According to
Martin, they are planning to sell many thousands of these systems. This support structure has been extended to
the enterprise space – they support Philips’ employees who use Apple computers
and iOS devices, for example.
As Martin says, “the dog got the bus” in terms of the unique space that
TruSource occupies. Unlike other support call centers that pay low wages and superficially
train, TruSource Labs pays front line support engineers $14 to $19 per hour vs.
the $9/hour that typical support call centers pay. TruSource’s higher wages coupled
with extensive training, said Alton, substantially reduces the typical turnover
that support call centers experience. Their attrition is 1/3rd the
industry average, which, of course, helps increase customer and end user satisfaction
all around while reducing costs. The idea, says Alton is to “pay people who do
the work more so overall costs go down.”
TruSource
Labs, whose COO, Cheryl Coppens, used to run Apple’s tech support in Sacramento,
follows the AppleCare and Genius Bar model of proving top drawer levels of
customer support for their customers. Many of TruSource’s staff are former
Apple support personnel. Alton talks about having “Apple’s support DNA” in
their systems, so that TruSource doesn’t have to teach a passion for customer
support, it’s already there.
According
to one of their customer, Rich Warwick, CEO of Plum, “While everyone else is
focused on trying to answer as many calls as possible with the fewest possible
low cost people in low cost regions by getting them on and off the phone as
quickly as possible, TruSource is focused on providing a truly great experience. While everyone else trains their support
teams to "make it someone else's problem as quickly in the call as
possible", TruSource trains their people to "solve the customer's
problem as fast as possible, regardless of whether it is our problem or
not".
Warwick
continued, “I think we have all experienced this on a call with Dell, HP or
Apple. Try connecting an HP printer to
an Apple computer then call either technical support with a problem and see how
quickly they tell you to call the other company. For this reason, TruSource provides core
training to all their employees for debugging generic networking and WiFi
issues, etc. Even if they do have to
send you to another company, they will gather the information the other company
needs to help you.”
TruSource
also offers a service called “Make it Better” which includes providing their
customers with non-geek speak marketing and support collaterals (which also
helps the agents in the call centers). They’ll do “deep dive analytics” on a
certain percentage of the customers who either call or email, so that TruSource
can alert product management to problems or suggestions as to how to improve
their product(s) or processes
TruSource
Labs can manage an entire beta product launch process, gathering detailed feedback
from beta users. This includes reaching out to beta users who dropped out to reengage
them in the process. With an industry
average of 20% drop out rate for beta users, this offers substantially more
data for product developers.
In
addition, TruSource monitors social media channels using sophisticated social
listening technologies. TruSource
monitors traditional channels such as Facebook, Twitter, Pinterest and many
others. Customer issues are immediately
addressed, and feedback provided to clients so they can address both product
and process issues.
Alton
noted many IoT companies have great products but very limited background in
support, and “it’s a huge education effort that we make with clients.” In
discussing clients’ needs, “just that level of conversation advances their
knowledge by several levels,” said Alton. “It’s meant a lot to them. We love transferring knowledge in order to
build strong teams and alignment between TruSource and our clients to make the
customer experience outstanding.”
They
not only want to do “good work,” said Alton, “but we want to work only with
companies that are brand builders, like Nest, who really care about their
customers, and with people we want to work with.”
As a former customer support fanatic, am delighted to see a support
organization, especially for IoT enterprises, that’s straight talking,
financially transparent and modeled after Apple’s excellent support “DNA.” Now,
if only Pebble would become a customer!
|
Photo of Alton Martin courtesy of TruSource Labs
© 2015 IoT Perspectives |
|
|