WHITE PAPER: How safe is your quantified self? from the Symantec Security Response Team

SECURITY RESPONSE
Fueled by technological advances and social factors, the
quantified self movement has experienced rapid growth.
How safe is your quantified self?
Mario Ballano Barcena,
Candid Wueest,
Hon Lau
Version 1.1 – August 11, 2014, 12:00 GMT

CONTENTS
OVERVIEW...................................................................... 3
What is quantified self?................................................. 5
What do self-trackers track?.................................... 6
Who actually does self-tracking?............................. 6
What can this data be used for?............................... 7
Case study: sports activity trackers......................... 7
How does it work?........................................................ 10
Types of tracking devices....................................... 11
Common self-tracking system models................... 13
Loss of privacy is a major concern............................... 16
Where are the risks?.................................................... 16
Data custodianship................................................. 16
Bring on the features, pile on the risks.................. 18
It’s personal data, but not as we’ve known it �� 18
Excessive information gathering............................ 19
What are the risks?...................................................... 20
Identity theft.......................................................... 20
Profiling.................................................................. 20
Locating of user or stalking................................... 21
Embarrassment and extortion............................... 21
Corporate use and misuse..................................... 21
The state of security in self-tracking........................... 23
Security issues seen in the field............................ 23
Where is self-tracking heading?................................... 29
Recommendations and mitigation............................... 30
For users................................................................. 30
For app developers and service providers �� 30
Conclusion.................................................................... 32
Appendix......................................................................34
Models of self-tracking systems.............................34
Resources..................................................................... 36

Fueled by technological advances and social factors, the quantified self movement has
experienced rapid growth. Quantified self, also known as self-tracking, aims to improve
lifestyle and achievements by measuring and analyzing key performance data across a
range of activities.
Symantec has found security risks in a large number of self-tracking devices and
applications. One of the most significant findings was that all of the wearable
activity-tracking devices examined, including those from leading brands,
are vulnerable to location tracking.
Our researchers built a number of scanning devices using Raspberry Pi mini computers and,
by taking them out to athletic events and busy public spaces, found that it was possible to
track individuals.
Symantec also found vulnerabilities in how personal data is stored and managed, such as
passwords being transmitted in clear text and poor session management. As we collect,
store, and share more data about ourselves, do we ever pause to consider the risks and
implications of sharing this additional data?
OVERVIEW

People are now
tracking every
facet of their lives
with the aid of
technology.
WHAT IS QUANTIFIED SELF?

What is quantified self?
In recent years the concept of collecting and analyzing data has moved from being mainly used in business to a
much more personal level. People are now tracking every facet of their lives with the aid of technology. This, in
essence, sums up what the quantified self movement is and what it stands for.
Today, self-tracking is big business and is experiencing rapid growth. A report by ABI Research estimated that
the number of wearable computing device shipments will reach 485 million units by 2018. The majority of these
devices will have tracking functionality. The number of wearable device shipments only accounts for tracking
devices and does not include smartphones that can run self-tracking apps, which would amount to billions.
According to a study by the Pew Research Center, 60 percent of Americans now regularly track their weight, diet
or exercise activity.
Whatever personal metric a person may choose to track, the goal usually boils down to trying to improve things
in some shape or form. You cannot better yourself if you cannot tell if you are better or worse than before. The
key to knowing where you are today is to measure and compare against past data, and that is the essence of the
quantified self movement.
The quantified self movement is now entering a golden age in its development because of a collision of several
forces at play in the world of technology, health, and popular culture. On the technology side, the ever-increasing
processing power and miniaturization of sensors and processors, improved battery life, and the rollout of
ubiquitous communications infrastructure has opened up a new world of possibilities for always-on devices that
can be carried around all day. Another key technology driver is the idea of big data and the wholesale collection
of personal data to gain insight into the behavior and habits of consumers.
In health, there is an increasing awareness among the public of healthier living. TV, radio, Internet, and print
media publications frequently promote health-related issues, products, services, and lifestyles. After many years
of bombardment about health issues, the message may finally be starting to sink in among the general public.
On the sociocultural side, there is a trend towards self-awareness, narcissism, and a need to publically express
personal opinions and views for social validation. The “selfie culture” and the rampant growth of social networks
are classic signs of this trend.
Figure 1. Factors driving growth in quantified self

What do self-trackers track?
Today, if you wish to track your activities, you are spoilt for choice when it comes to apps and devices that can
help. A quick trawl on the Web reveals that there are apps available to track a multitude of subjects:
All that a tracking app or device does is take readings of states from various sensors, then digitizes and stores
them for future use. If there is no current or practical sensor technology available to track a certain subject, the
infinitely flexible app UI can allow the user to make an assessment of the subject’s current state and input data
into the app. An example of this is for mood tracking or water consumption.
While much of the information listed is not that sensitive on its own, some of the data could be considered highly
sensitive. For example, while medical data requires careful handling, the amount of water you drank yesterday
would not generally be considered sensitive information.
Who actually does self-tracking?
Regular practitioners of self-tracking include people with chronic medical conditions who track their symptoms
to try and establish patterns in their state, which could help identify correlation factors for their conditions. They
may do this as part of a medical care regime or just out of personal interest.
Another common type of self-tracking practitioner is the sports enthusiast. A keen runner could collect
data about their running activity to help them set performance goals and track progress. By keeping a log of
performance data, a sport enthusiast could determine whether they are improving or not.
Aside from these two types of users, there is a broad swathe of other people who may be just curious or wish to
achieve something, such as giving up smoking, losing weight, getting more sleep, or living a generally healthier
lifestyle. While the health benefits of many self-tracking devices and apps cannot be scientifically proven, many
people clearly believe they are beneficial, as the growth figures in self-tracking apps and devices show.
There are also self-tracking geeks who are interested in documenting all facets of their daily lives in as much
detail as possible in public and have turned the whole idea into an art form. Perhaps the most extreme example
is Alberto Frigo, who has embarked on an extended journey of discovery to track every detail of his life for
36 years. He aims to record, collect, and photograph a gigantic number of aspects about himself and his
environment. His aim is to create a comprehensive record of his life and experiences. Frigo started his journey
in 2004 and is currently 10 years into the project. So far, he has recorded a mind-boggling 295,000 photos of
things that he has used or interacted with, over 12,000 dreams, over 600 photos of new acquaintances, 7,500
Table 1. Examples of types of information that can be tracked using self-tracking apps
Consumption
• Calories/food
• Alcohol
• Nicotine
• Caffeine
• Water
• Drugs/medicine
Bodily functions
• Body PH
• Menstruation/Fertility
• Pregnancy
• Stool/bowel motion
Physical activity
• Sports activity
• Sleep
• Travel
• Sexual activity
• Tooth brushing
Medical symptoms
• Headaches
• Pains
• Asthma attacks
• Allergies
Spatial
• Location
• Altitude
• Time
• What you see
Physiological statistics
• Heart rate
• Blood sugar/glucose
• Temperature
• Blood pressure
• Weight
• Breathing
Mental health
• Mood
• Stress levels
• Alertness

drawings of ideas, and 285 square meters of trash collected from his walks. These are just some of the things
that he has collected and recorded so far.
What can this data be used for?
Aside from the clearly stated self-improvement use case, some of this new type of information could also be
extremely useful to marketers. Marketing is all about finding out what people want and offering it to them at
the right time. Self-tracking data is potentially a goldmine for marketers because it can allow them to gain
deep insight into an individual. For example, let’s say you like running so you collect and track all of your
running activities and upload them to the service provider’s cloud servers. By accessing this data, a sports shoe
manufacturer’s marketing team could learn a lot about your running habits, such as:
• The mileage that you are covering
• When you usually go running
• Where you usually go running
• Where you live
• Your age, sex, height, and weight
• Where and when you are on vacation
Based on this data, marketers could derive valuable insight into your habits and behavior, and could target
marketing campaigns to you such as:
• Sending offers for new running shoes when you are nearing the typical shoe replacement mileage
(300-500 miles).
• Sending offers for the right type of running shoes. For example, they could offer trail running shoes if the GPS
data indicates that most of the user’s running activities take place on trails rather than roads.
• Adjusting the price of products and services based on a user’s location.
• Sending offers from retail outlets that you often pass by.
All of this knowledge can be inferred either by analyzing data that you typically provide when you sign up or by
reading data that is generated during the use of the device or service. However, the really powerful use cases
happen when this information is combined with data that has been gathered from other sources. This gives a
much more complete picture of the person in question, allowing for far more accurately targeted marketing.
Case study: sports activity trackers
One of the first types of quantified self applications that gained wide user acceptance is the now almost
ubiquitous sports activity tracker. This genre of application took off a few years ago when GPS trackers became
small enough to be integrated into watches and mobile phones, allowing users to easily carry the trackers
around. Using these applications, users can track their sports activity, such as a running session. The tracked
data may include start and end times, speed, current location, the route taken, the altitude and so forth. Often,
the device that performs the tracking is married to an online service where the data is uploaded. Once uploaded,
the data can then be analyzed.
To try and picture the amount of data that could be collected in just one session, let’s assume the sports app
collects the following data at regular intervals during a session:
• GPS location
• Time
• Heart rate
• Speed
• Altitude
• Steps taken

Data readings are taken at short intervals to provide reasonable accuracy. Suppose the data was sampled once
every ten seconds during the session. A one hour running session would generate 360 readings. Suppose the
service had one million users and they all used it for an hour every day. That’s 3.6 million sets of data generated
in one day by just one app. It actually could be quite conservative to assume that a sports tracker app has one
million users. A quick scan of a number of popular self-tracking Android apps and services reveal the following
download numbers:
• Runkeeper – 10 to 50 million users
• Runtastic – 5 to 10 million users
• MapMyRun – 1 to 5 million users
• Strava Cycling – 1 to 5 million users
• Fitbit – 1 to 5 million users
• Jaw Bone Up – 1 to 5 million users
Based on these numbers, it is clear to see that there is potentially a lot of data being collected, transmitted,
and stored on various servers around the world. In today’s world, where information is the real currency,
these servers are potential goldmines ripe for exploitation. The information could be useful to governments,
marketers, businesses, and of course cybercriminals such as the Cyclosa gang who were behind the SSNDOB
attacks.
For example, one of the services we looked at states the following key selling points for their device:
• Allow you to discover previously unseen patterns in activity and gain insight into your daily life.
(You aren’t the only one who might like to know this)
• Track your sleep patterns so you can know when you have the deepest and lightest sleep.
(This might be useful for your local burglar to pick a good time to break in)
While the sports tracker usage scenario is relatively benign, there are some instances where leaked self-tracking
data could potentially be more damaging or embarrassing if it ended up in the wrong hands.

While some
die-hards still use
pen and paper,
most self-tracking
today is done
using an electronic
device.
HOW DOES IT WORK?

How does it work?
Self-tracking is all about turning everyday activities, thoughts, and statuses into discrete data that can be stored,
analyzed, and then used to guide a process of change which will hopefully lead to a desired outcome.
The process of self-tracking operates like a cycle and typically works like this.
1. Track and collect data from an activity
2. Analyze and compare performance and status against a desired goal
3. Make adjustments based on findings
4. Repeat process
While some die-hards still use pen and paper, most self-tracking today is done using an electronic device.
After data is collected, it needs to be analyzed using software. The knowledge gained from the analysis is then
fed back into the process to help guide the user towards a goal. We will now take a closer look at the common
classes of self-tracking devices.
Figure 2. Quantified self cycle: Track, analyze, and adjust

Types of tracking devices
There are many ways to self-track activities. An increasing number of people are carrying smartphones and
devices with them all day every day, which can be used to collect data. Some tracking devices can also be used to
review and analyze data. Devices used in the domain of the quantified self typically come in one of two guises: a
smartphone or a wearable device.
Smartphones (with apps)
Most modern smartphones have a plethora of sensors built into them and many of these have self-tracking
applications. Smartphones are primarily telephony devices but the inclusion of multiple sensors along with a
suitable app execution environment can turn them into general purpose, self-tracking devices. Built-in sensors
may include an accelerometer, gyroscope, barometer, heart rate sensor, thermometer, proximity meter, ambient
light sensor (light level, usually the camera), and navigation systems such as a digital compass, GPS, and
GLONASS.
What these devices actually track depend on the apps the user installs. The app provides the framework to let
the device read, store and interpret the signals generated by these sensors. Apps can also let users track data
Figure 3. Typical sensors found in modern smartphones that can be used for self-tracking purposes

that sensors cannot currently capture, such as data on moods, food and drink consumption, aches and pains,
etc.
Wearable tracking devices
Wearable devices are designed to be worn on the body. These devices typically have a small and light form
factor, letting users wear them on the wrist like a wristband or a watch. Alternatively, they can be attached to
sports equipment such as running shoes, clothes, bikes, etc. These devices usually contain accelerometers and
gyroscopic sensors and these sensors are responsible for generating the data. By reading the stream of data
from these sensors and then applying data processing algorithms, the devices can recognize patterns to identify
the wearer’s current activity.
Most current wearable tracking devices have a limited user interface, such as a single touch point and a few
LEDs indicating the device status or a small dot matrix display. However, this is changing with the increasing
number of new smartwatches. Today, the vast majority of wearable devices only perform a data collection
function. They require a separate computing device to let the user access data analysis functions.
Figure 4. Typical wearable activity-tracking device

Common self-tracking system models
Many self-tracking devices, particularly the wearable type, do not have a suitable user interface for analysis and
data reviewing. Instead, data must be transferred to another place where it can be aggregated and processed
before being presented to the user in an easy-to-understand format.
There are generally three pillars to quantified self systems. Each system may use one, or more commonly, a
combination of the pillars to provide an overall service to the user.
There are a few general models for self-tracking systems, and they reflect where the processing and storage of
data resides. Essentially, data is collected either by a wearable device or a smartphone. The data may then be
stored, processed, and presented by the wearable tracking device, the smartphone, or the cloud service, or a
combination of the three.
We will now look at two of the most common system models used in quantified self systems. There are other
system models used in self-tracking which are detailed in the appendix.
Smartphone app + cloud
Building on the previous system model, in this approach, the smartphone incorporates the required sensors and
is used to collect, store, analyze, and present the data. It extends the functionality available locally on the phone
by allowing the locally collected data to
be synced to the cloud, where additional
advanced analytics and services can be
offered, possibly at an additional cost by
way of a subscription.
This is the most common system model for
self-tracking, perhaps for several reasons:
• There is a massive indie app
development movement. In general, it
is easier to build software apps than
hardware. Building hardware requires
skills that are harder to find and will
require considerably higher investment
in terms of money and labor. It is
much easier, particularly for smaller
companies, to just build software only,
rather than building software and
hardware.
• There is a large pool of existing
smartphones already deployed and
Figure 6. Most smartphone-based self-tracking apps sync with a
cloud-based service
Figure 5. Three pillars of quantified self - tracking devices, smartphones and computers, and cloud servers

actively used. Developers can
quickly and easily use the existing
rich execution environment and
development toolkits to build apps.
• Smartphones already contain many
of the required sensors for most
self-tracking applications. Since
they are already built into the
phones and available for use, users
do not need to spend more money
on extra hardware, making this
option more attractive for users.
• Users simply prefer not to carry
multiple devices. It is much more
convenient to just carry a single
compact device that can perform
multiple functions than to carry
multiple single-function devices.
This is true as long as the general
purpose device (smartphone) offers
a comparable result to specialist
hardware. For many functions, this
is already the case. For example, a
smartphone based app can do an
equal or better job of tracking GPS
coordinates and monitoring trip
performance than many specialist
hardware-based solutions.
Figure 7. Examples of smartphone-based quantified self tracking
apps that use cloud-based services

Tracking device + smartphone app + cloud
This model is relatively common in systems that use wearable tracking devices. One reason for this may be to
reduce costs by leaving out the network adaptor hardware from the device, instead leveraging the one found
on the smartphone. Other reasons include keeping the weight and size of the devices down and preventing
battery drain. For wearable devices, the size, weight, and battery life are critical factors for its usability and
comfort. A wearable tracking device is responsible for providing sensor readings and transmitting the data to the
smartphone. Data is typically synced to a smartphone using a wireless mechanism such as Bluetooth Low Energy
or ANT+, but sometimes a physical connection may be used. The smartphone app is then periodically synced to
the cloud through an Internet connection.
In this model, the smartphone stores some or all of the recorded data, but a copy is also sent to the cloud servers
for storage and more detailed analysis as well as other value-adding features. The latter is becoming a more
common way for service providers to monetize the collected data by making more detailed or advanced analytics
available at an additional price to the user.
Note: A smartphone in this context could also be a desktop or laptop computer with Internet access.
Figure 8. Self-tracking system model using a device, smartphone, and cloud-based server

Loss of privacy is a major concern
According to the results of a survey conducted by Pew Research and the Carnegie Mellon University published
in September 2013, 86 percent of adult Internet users in the US occasionally took some steps to try and protect
their privacy when online. This figure shows that a significant number of users have an awareness of the risks to
their privacy when using the Internet. It also shows that these users are concerned enough about the problem
to take some steps to avoid being tracked or monitored when online. Some of the steps taken by users included
encrypting communications, using anonymity services such as proxies, Tor, or VPNs, and using a fake name or
giving inaccurate information about themselves. The growth of the CryptoParty movement, whose mission is to
educate the public on how to use encryption and privacy tools, is a sure sign of a growing public unease about
organizations and governments snooping on user’s online activities and habits.
Another interesting finding of the survey is that users were most concerned about having their online activities
seen by criminals and advertisers. Concern about being monitored by governments was almost at the bottom of
the list.
Given this background, it is interesting to note that perhaps the greatest overall risk posed to users by the
quantified self movement is the risk of the loss of privacy. Never before has such a huge amount of information
been collected, transmitted, and stored about users. People are freely and actively engaging in the collection of
information about themselves and, as we will see, the risk to privacy is not improving.
Where are the risks?
When it comes to self-tracking, there are several ways to collect, store, analyze, and present data. Many services
involve multipart systems but unfortunately, as more parts are introduced into a system, more risks are added
to the equation. Each extra layer in the system increases the risk of attack, as the new elements introduce new
potential weaknesses and points of failure which could be exploited by attackers.
Data is generally at risk either at rest or during transmission. What we mean by at rest is when the data is stored
for archival purposes such as in a database. Databases can be local, remote, or both. During transmission is
when the data is being sent from one device or location to another. Transmission could be performed locally and
offline or it could be remotely and online. It can also be done in batch mode or continuously.
Data custodianship
The lifecycle of data handled by most self-tracking systems involve three stages. There is a local data collection
phase, a transmission phase, and a cloud-based storage and analysis phase with potential feedback loop. Given
this setup, there are three main risk areas for the data collected by self-tracking apps:
• On the device (storage)
• In transit (transmission)
• In the cloud (storage)
On-device risks
Scope of risk: Data about a single user
Data stored on the device is generally about a single user as these devices are usually for personal use. The data
stored locally is at risk from malware that can steal data locally. Symantec observed that information stealing
is one of the most common traits of mobile malware in 2013, accounting for 28 percent of the threats and 30
percent of mobile malware tracked users. If criminals find valuable data on a device, they are inevitably going
to target it. To mitigate this risk, you need proper access control and permissions on the data. The sandboxing

of data is built in to Android and iOS to prevent one app from seeing and interfering with data from another
app. This works for the most part, as long as the device is not rooted and no vulnerabilities are found that can
circumvent these controls. Encryption of locally stored data should also be considered if the data is considered
sensitive enough.
Another obvious risk to locally stored data is the threat posed by the theft of the device. Many self-tracking
devices do not offer much in the way of protection in case of physical theft. On smartphones, users can at least
make use of the phone locking feature to prevent unauthorized access to data, should the device be stolen.
Transmission risks
Scope of risk: Data about a single user or limited number of users
Data collected by self-tracking apps and devices often need to be sent to the cloud either in real time or in
batches, such as at the end of an activity session. Transmission may occur directly from the device to the cloud
or from the device, to a computer, and then to the cloud. Indirect syncing may involve the use of short range
radio technologies such as Wi-Fi, Bluetooth, or NFC, or cable-based syncing. All of these methods have their own
security issues to deal with.
During transmission, data is at risk from an array of possible threats. These include traffic sniffing, which
lets attackers collect all transmitted data, and man-in-the-middle and redirection attacks, which could cause
data to be sent to the wrong server. One way to mitigate some of these risks is to apply strong encryption and
authentication on the data being transmitted. With Wi-Fi, for example, the link could be encrypted with WPAv2.
For the local-to-cloud leg of the connection, a network-level security solution such as TLS and a VPN should be
used on untrusted networks. Depending on the sensitivity of the data being sent, the data may also be encrypted
at the application layer.
Cloud storage risks
Scope of risk: Data about all users
Once data arrives at the cloud destination, it is processed, collated, and stored in a central database of some
shape or form. The fact that the database can receive data from remote app clients means it is exposed to the
outside world to a lesser or greater extent. This exposure means there is a risk of compromise. Depending on the
configuration of the system, there could be any number of risks including SQL injection attacks, account brute-
force login attacks, distributed denial-of-service (DDoS) attacks, remote software vulnerability attacks, default
password or back door attacks.
The risk in the cloud has a much wider scope. An attacker could break into a single user account or they could
compromise the whole system and every user account stored in it by targeting the systems of the service
provider or by targeting its staff. Over the years, we have seen countless mega data breaches across a whole
variety of industry sectors including healthcare, hospitality, retail, industrial, defense, and government. There is
no reason to believe that cybercriminals would be less interested in quantified self data, particularly if it is co-
located with other personally identifiable information (PII) such as social security numbers (SSNs) and payment
card data.
What can be done? The use of good access controls, strong passwords, and solutions like two-factor
authentication (2FA) could help prevent account compromises. Because cloud service providers must
expose their service interfaces to the world at large, they are vulnerable to probing and targeted attacks by
cybercriminals who wish to gain unauthorized access to the data. Consequently, service providers have a major
challenge to ensure that their systems are built and provisioned securely, and are adequately protected on an
ongoing basis.
How self-tracking data is managed in the cloud is generally outside of the control or visibility of the users,
but they could still get clues as to whether the service providers are handling data in a diligent manner. Users
should look for privacy and security statements (for example iCloud, Fitbit, and Jawbone) and compliance with
standards such as PCI-DSS, HIPAA, or ISO 27001 where appropriate.

Service providers should at least encrypt all data whether in transit or at rest (you can never be too safe in this
age of mega breaches), and there should be appropriate access controls to the data – DLP solutions could help
prevent unauthorized access and copying of data. Data should be appropriately segregated too; one user should
never be able to access another users’ data. Service providers should also consider anonymizing user data as
an additional, but not foolproof, layer of security. For example, having a set of GPS coordinates that cannot be
linked to a person or time makes the data less useful to attackers.
Bring on the features, pile on the risks
Arguably, the more features and functions that are added to a system, the more complex it becomes and,
consequently, the chances of it being less secure increases.
Take, for example, a standalone activity-tracking device with its own data storage and display for showing data.
On its own, it does not pose much of a privacy risk. The only risk of anybody else finding out what the owner has
been doing is if the attacker managed to gain physical access to the device.
Suppose we add a wireless syncing feature to the device, allowing it to sync to a smartphone app with a better
display and more data storage. Syncing in this case is done using Bluetooth Low Energy, a short range wireless
communications protocol. By adding this feature, not only have we added more options for usability and
functionality, but we have now introduced the risk that an attacker could remotely sniff the data that is being
sent over the airwaves. Alternatively, attackers could undermine the wireless device syncing mechanism by
trying to hack into it through security weaknesses or by forcing or tricking the device to connect with a computer
controlled by the attacker.
Even if data is synced by wire, merely storing the tracking data on another device increases the risk profile,
because now the data is stored in two places. In many cases, the device where the data is synced to tends to hold
both personal and aggregate data. The volume of data is much greater and increases the potential impact if an
attacker compromises the data.
Suppose we now extend the functionality further by adding online cloud service functionality. After a user has
synced their data to the smartphone app (which is still within the user’s physical domain), they now have the
option of uploading their data to an online cloud service for safekeeping, analysis, and social sharing. Add social
media integration and an API to allow third party developers to build apps that leverage the data and the risks
mount up.
Unlike data in the user’s domain, the cloud service domain is mostly outside of the users’ control. Users have
limited control over authentication, authorization, access, and sharing. Almost everything else, including the
responsibility for control and security, is handed over to the service provider who chooses how to protect the
data, how it will use the data, and who it will share the data with.
With the addition of a cloud service layer into the system, suddenly the attack surface is much larger and
more difficult to defend. The risk of attacks can now come from remote locations and attackers can attempt to
intercept network traffic to steal data from individual users or target the mother lode by attacking the cloud
service provider directly. A successful compromise of a cloud service provider, allowing access to the user
database, could compromise all of the users of the service.
It’s personal data, but not as we’ve known it
Over the years, many of us have grown accustomed to sharing a certain level of information with online service
providers. For example, many services request a full name, date of birth, phone number, address, email address,
password, and security questions and answers when the user signs up. Most of us have come to accept this as
a trade-off for using the services. Some information, such as the IP address, may also be involuntarily collected
and recorded without any user intervention.
The risk to privacy increases considerably as the amount and range of data known about us increases. When
it comes to risk, everything depends on the context. A single piece of personal information on its own, such as
a date of birth, provides no context and poses few risks to anybody. Thousands of users in a website database

could share the same date of birth, so that information on its own does not allow any one person to be singled
out. However, if we can associate a first and last name to the date of birth, suddenly the number of matching
users in a database may shrink dramatically.
When we talk of “personal information,” names, contact details, and dates of birth are the type of information
that we typically think of. We could call this traditional personally identifiable information (TPII). But now, new
technologies enable us to collect much more information at a deeper and more personal level. Data generated by
self-tracking devices and services (also known as first-party data) is potentially highly personal and could reveal
an awful lot more about ourselves to others than we may like.
Traditional PII can tell somebody about who we are, where we live, and how to contact us. Additional information
generated by self-tracking services can tell somebody about what we do, where we are or have been, and when
and potentially why we are doing something. When additional self-tracking information is combined with
traditional PII, the potential for abuse becomes even greater. As more data is aggregated and relationships
between data are formed, data becomes information and, after further analysis, becomes insight which can be
used to predict the future behavior of people. This is gold dust to marketers as well as cybercriminals.
Excessive information gathering
While it is understandable that some self-tracking services need to know certain personal information in order
to provide a useful service, some services ask for a lot more information than is really necessary. For example, it
would be reasonable to expect a sports activity tracking app to ask for the user’s age, gender, weight, and height
in order to calculate certain statistics such as the estimated calories burnt during an exercise routine. However,
would such a service need to know the user’s home address, relationship status, education, or work background?
Would these factors have any bearing or impact on the user’s fitness or the provision of the service to an
Figure 9. Traditional PII versus self-tracking information - tells a different story

individual? Some of these data points are more appropriate for a consumer survey so before signing up, users
need to ask themselves whether they are signing up for a service or a marketing survey.
There’s a saying that goes “If You’re Not Paying For It, You Become The Product.” Many online services offer their
products for free but these are businesses and they have to generate revenue to survive. Instead of making users
pay upfront, they use revenue generation strategies such as advertising and collecting and selling marketing
information. This is a trade-off that users need to be aware of when using online services.
The next time you are asked for your date of birth, ask yourself whether the date of birth is really needed or if a
year of birth would suffice.
Thankfully, a lot of this extra information gathering is optional, but users should be aware of what is a
reasonable amount of information required to effectively operate a service compared to unnecessary and
excessive information gathering.
What are the risks?
The following are some of the things that your private self-tracking data could potentially be used for:
Identity theft
There is a criminal industry that prospers on gathering and selling as much PII as they can get their hands on.
Complete sets of data about a person can then be sold to other criminals in packages known as “fullz.” In this
criminal business, the more complete and up to date the set of details is, the more valuable it is. Having a more
complete set of data about a person can allow fraudsters to better fake official documents or commit other
frauds. For example, details could be used to set up false bank accounts for money laundering, ransom attempts,
or IRS fraud through fraudulent tax returns.
The threat of data theft or misuse does not have to emanate from outside of an organization. There have been
many incidents in the past of rogue employees selling customer information to third parties for personal gain.
These range from small scale incidents to massive data theft incidents involving millions of users.
Profiling
Many organizations already use profiling to target, exclude, or even discriminate against certain types of people
based on personal information that they have collected about them. Details provided by users to self-tracking
services could enable marketers and government agencies to organize and target certain types of users. Profiling
is of concern to privacy and human rights advocates because it can be easily misused to the disadvantage of
certain groups or minorities.
Insurance is likely to be one of the key beneficiaries of self-tracking data. There is already much discussion about
how insurance could use quantified self data. We are already seeing some limited use of self-tracking data for
insurance applications where some employers are using it to discount health insurance policy premiums based
on certain perceived positive behaviors such as leading a healthy lifestyle.
In usage-based or telematics insurance, tracking devices (which can be considered as self-tracking devices) are
fitted to cars to track the driving habits and performance of drivers in return for reduced insurance premiums.
More time on the road combined with bad driving habits such as hard acceleration, late braking, and fast
cornering - behaviors that fit the profile of aggressive driving - may result not only in an increased fuel bill and
added wear and tear to the car but insurance premiums will likely increase too.

Locating of user or stalking
Location-based self-tracking PII could also be abused for criminal purposes if it was to fall into the wrong hands.
For example, if criminals were to gain access to a sports tracking database, they could determine where a person
lived and when they would be most likely to be away from home and plan a break-in accordingly. Over the years,
there have been stories of people having their homes broken into while on vacation due to social media postings
of pictures and updates showing that they were away from home. You could obtain this kind of information from
many self-tracking service cloud databases.
Accurate and real-time location-based tracking can be useful for some activities, but this information could also
be useful for stalkers and even private investigators and governments who can use this information to locate
their targets.
In some countries, the police can use average speed detection systems to catch speeding drivers. These systems
work by capturing the time and license plate of a car as it enters a road segment and then again when the car
leaves the area and works out the average speed to cover the distance between the points. This system would
not be necessary if everybody had tracking devices in their cars and the police had access to monitor the location
and movements of all vehicles in real time, but the privacy implications of this level of mass surveillance might
be too much for it to ever come to pass.
These are just some of the risks that can arise from having your location tracked and exposed. There are
countless other scenarios where you may not want your location to be known.
Embarrassment and extortion
We have already seen that self-tracking services can track much more than just fitness and sports performance.
There are many services that track medical or health related activities and bodily functions. For example there
are self-tracking applications that can track the mood, toilet, and sexual activity of users. Are these the types of
data that we would be happy to share about ourselves? Does anybody but yourself, or perhaps your doctor, need
to know when the last time you went to the toilet was? Would you be comfortable with the exposure of this type
of information about yourself to the world? Perhaps a person who suspects that their partner may be cheating
might be interested in data collected from these devices.
We have already seen many cases of sensitive information (intimate photos or videos for example) falling into
the wrong hands and then being used to extort money from victims. As more and more sensitive information
is collected and transmitted around the world, the risk of highly sensitive data falling into the wrong hands
increases.
Corporate use and misuse
Self-tracking service providers have not been slow to catch on to other potential business applications for data
generated by self-tracking technologies. Some vendors in activity tracking technology actively promote the use
of their devices in corporate wellness programs. The benefits touted may include cheaper health insurance,
decreased sick leave, decreased healthcare costs, and even increased productivity due to employees feeling
more positive as a result of a more active lifestyle promoted by use of these types of devices.
Not all corporate uses garner a positive reaction. A supermarket chain in the UK uses tracking devices to monitor
the activities of staff to help improve work efficiency. The staff members who are required to use these devices
have expressed mixed views about them, with some saying that these devices are putting staff under immense
pressure to perform.
Despite many businesses’ best intentions to keep customer information safe, their databases are often the
target of cybercriminals who attempt to break in and steal the information. There are also insider threats
to businesses and there have been many stories in the past of employees of businesses stealing customer
information and passing it on to various agencies for a fee. Insiders can pose a considerable threat as staff may
have privileged access to data, making it easier for them to copy or manipulate it.

Users of wearable
self-tracking
devices can be
tracked remotely
without them
realizing.
THE STATE OF SECURITY
IN SELF-TRACKING

The state of security in self-tracking
Self-tracking apps and services are essentially just another category of apps that run on mobile devices. What is
common about them all is that they are designed to record information about the person. Because these services
are just another type of mobile app, it should come as no surprise that they share many of the same kind of
security issues that have been seen in other mobile apps.
For research purposes, we examined a range of popular self-tracking devices and mobile phone-based health
and fitness apps to see what kind of security issues we could find.
Security issues seen in the field
Granular location and personal tracking
KEY FINDING: All of the wearable activity tracking devices examined, including those from leading brands, are
vulnerable to location tracking.
This is tracking the location of a person to a high level of granularity and detail. In some self-tracking
applications, there is a necessary and expected trade-off between privacy and functionality. This scenario
typically applies to GPS-based activity trackers such as smartphone sports tracker apps or GPS-based self-
tracking devices like sports watches. People use these devices and services knowing full well that they will be
tracked and are happy to do so because they choose to have their location tracked during the session.
What is more problematic is when users are tracked without knowing they are being tracked or can be tracked.
This can arise even when people are using devices that have no obvious means for location tracking.
Most self-tracking devices are simple in their makeup. They typically contain a few physical sensors such as
accelerometers, gyroscopic sensors, and tilt meters. These are used to detect patterns of motion that are then
Figure 10. Many Bluetooth-enabled devices can be easily tracked

interpreted as certain activities. In our research, we found that all of the devices we looked at either had a USB
and/or a Bluetooth interface, most commonly Bluetooth Low Energy (aka Bluetooth Smart). The latter option
allows the self-tracking device to be connected part-time or continuously to another computing device without
wires, which makes it very convenient for users as well as those who want to track them. As it turns out, many
of the current activity-tracking devices on the market, such as the sports wristbands or pendants, can be easily
tracked.
The key to how these seemingly offline devices can be tracked comes down to how they use the Bluetooth LE
technology. Many Bluetooth LE devices are assigned a specific hardware address, much like the way standard
computer network cards have a fixed MAC address. When in use, Bluetooth LE-enabled devices can transmit a
short range (<100m) signal that advertises itself to nearby devices. This signal can be read by anyone within
range and, depending on the device, it may contain information necessary for a connection to be established.
This information also includes the fixed Bluetooth LE network address. In some cases, the devices may also
expose other information that could be used for tracking the device. The information may include serial numbers
or other internal IDs specific to the device and could be accessed simply by performing a remote querying
operation on the device.
By placing a number of scanning devices at various locations, it is possible to scan and locate a device. By
identifying the hardware address and measuring the relative signal strengths between scanners and the device,
it is possible to get an approximate fix on the physical location of the device.
For this project, we built a number of Bluetooth scanning devices using Raspberry Pi mini computers coupled
with a battery pack, an SD card and Bluetooth adapter, for the low price of US$75 each. (We call them Blueberry
Pi!) We placed them at selected points along the course of a major European running event to see what we would
find. When we looked at the data, we found that some of the runners were self-tracking devices that could be
tracked using Bluetooth scanners. By placing scanners at various points of the race, we could determine when a
Figure 11. One of our Blueberry Pi scanning devices

device passed the scanners and, by comparing times when the device was picked up by different scanners, work
out the average speed of the competitor.
We also performed a similar scanning exercise by walking through the busy city center streets of Dublin, the
capital of Ireland, and also at public transport hubs in Zurich, Switzerland. During these separate scans, we also
picked up a range of self-tracking devices and a large number of mobile phones and tablet devices. These simple
scan results show that it is relatively easy for these devices and, by extension, their owners to be tracked.
The scans also showed that there are quite a few people wearing these devices when going about their business
around the city. Interestingly, during our scans, we found that one vendor for self-tracking devices dominated
the marketplace, with three quarters of devices found coming from this one vendor.
Tracking problems are not limited to Bluetooth LE devices. The problem has been well documented in Wi-Fi.
Currently, as you walk around with Wi-Fi enabled on your portable device , your device is giving away its unique
MAC address, which means it is possible for somebody to track your device (and¬ by extension – you) from afar.
In recognition of this issue, Apple recently announced that one of the changes to be included in iOS 8 is the use
of randomized network IDs when the phone is scanning for Wi-Fi access points. Only when the user chooses to
connect with a found Wi-Fi network will the true MAC address be revealed. This shows that major vendors have
recognized that network address tracking and its privacy implications are a real threat to users.
Transmission of tracking and personal data in clear text
KEY FINDING: 20 percent of apps transmitted passwords in the clear
Most of the self-tracking services that we looked at required or offered online cloud-based service components
for which users have to create an account for in order to use. Whenever there are user accounts, user names
and passwords are never far away so we were interested to see how the different services handled sensitive
information such as login credentials. We were disappointed to find that out of all the apps that we looked
at that required user logins, 20 percent of them transmitted user login credentials in clear text, meaning no
attempt is made to encrypt the passwords at all. In a couple of cases, unsalted MD5 hashes of the passwords
were sent, perhaps as a way of securing the passwords, but unsalted MD5 hashes are easily crackable with
rainbow tables so this offers little protection. In some cases, due to deficiencies in the design, a cybercriminal
could just use the password hash itself (no need to crack the hash) to log into an app, as the hashing of the
submitted password is performed in the app (client side).
What is particularly worrying about this finding is that there is already ample evidence available to show that
many people reuse the same user name and password for multiple services. It only takes the weakest link in the
chain to expose credentials which could then be used by attackers to take over other accounts that have a more
secure setup.
In other cases, self-tracking and other personal data may also be transmitted without the use of a secure
channel. Transmission of data in clear text leaves the user data wide open to data sniffing. The classic example
is when users are connected to unsecure Wi-Fi networks to transmit data. A number of apps that we looked at
exposed the email address and other account data as well as details of the user’s activities in this way. Given
that so many of the apps transmitted login credentials in clear text, it does make us wonder how the data that
is stored on the server side of these services is treated. We have seen plenty of cases of server data breaches
where user names and passwords were stored in plain text in databases. The bottom line is user credentials and
data should always be encrypted at rest and during transmission and vendors need to ensure that they handle
sensitive data appropriately.

Lack of privacy policies
KEY FINDING: 52 percent of apps examined did not make available privacy policies
Self-tracking apps and services are by their nature designed to collect and analyze personal information.
Therefore it is reasonable to expect and indeed is legally required (such as in the Online Privacy Protection Act
2003) of companies that collect and manage PII to make a privacy policy available that is displayed prominently
and easily accessible. Privacy policies should preferably be understandable even by those not in the legal
profession and must be shown to users before they sign up for a service so that they can make a considered
choice before using it. The policies should typically explain to the user the following:
• Who is collecting the data?
• What is being collected?
• When is data collected?
• What will the data be used for?
• How long will the data be kept?
• How can the user access and control the data?
• Will the data be shared with third parties?
Privacy policies are important because they form part of the user contract. While privacy legislation may vary
from country to country, in general, information that is collected may only be used for the originally stated
purpose.
Despite the importance of having a privacy policy, the majority of apps did not have one. Of the other 48 percent
that did have privacy policies, many of them used generic privacy statements with vague promises of keeping
user data private without any elaboration. The lack of a privacy policy may be a possible indicator of how the
issue of security is treated in the development and provision of online self-tracking services. Users would be well
advised to take this into consideration before signing up for any services.
Contacting multiple domains
KEY FINDING: The maximum number of unique domains contacted by a single app was 14
The average number of unique domains contacted by the self-tracking apps that we looked at was five and the
maximum number was 14. While it is understandable that apps may need to contact a few domains in order to
transmit collected data and access certain APIs such as for ads, it come as a bit of a surprise that a significant
number of apps contacted 10 or more different domains. The types of domains contacted can be categorized
into the following areas:
• Service provider (to transmit user data)
• OS provider
• Ad networks (Tapjoy, Doubleclick, Amobee, Simpli.fi)
• CRM/Marketing services (Apsalar, Localytics, Apptentive, Flurry, Admob, Appsflyer, Aro, Uservoice, BudURL,
Mixpanel, Adjust, Kiip, Urbanairship, Fiksu, Google Analytics)
• App analytics and testing (Crashlytics, Crittercism, Testflightapp, Bugsense, Newrelic,)
• Social media APIs
• Utility API (Forecast, Wunderground, Appspot, Mapping services)
• App frameworks (Parse, Amazon Web Services, Appspot)
While the apps may have legitimate business reasons for contacting many different domains, from a user’s
perspective, many of the domains being contacted are receiving information on the user’s behavior and activities
(metadata) without the user actually being explicitly informed about it. In some cases, a reference may be made
in a privacy policy but in general, metadata collection appears to be considered as fair game. Therefore, it is not
surprising to find that many of the third-party domains being contacted belong to CRM/analytics and marketing
services. These services allow the app provider to monitor and track user behavior in relation to how users use
the app and respond to different offers and features. It is great for the app developers because it allows them
to conduct user research and gain insight into user behavior, but it is not so great for using the users’ data plan

allowances and their privacy. This point neatly leads us onto the next issue of unintentional data leakage.
Weak session management and security
With user account-based services, one of the dangers is that session management in the service may be insecure.
Weaknesses in this area may enable attackers to guess user accounts in the system and then hijack sessions or
access data belonging to other users in the system. One of the simplest ways to do this is to guess or even simply
increment the session or user ID while logged in with a valid session of another user. Poorly designed systems will
permit this to happen and reveal data from other users. In our study, we found a number of apps that showed this
type of weakness. One particular system was so poorly designed that it could expose user accounts data if you know
the email address of one the users of the system or if you simply modified the user ID in the request as the IDs are
sequential.
Session IDs should be large alphanumeric strings that are randomly generated by cryptographically safe methods
and should only be valid for one given user ID. Clearly some vendors are falling well short of the mark.
In one case, we found a particular app that exchanged whole SQL statements with the server to create new tables
and update them. This type of setup could open up a huge security hole that can be exploited by attackers to gain
unauthorized access or manipulate the database by modifying queries sent by the app.
Unintentional data leakage
Despite the best intentions of app developers, information about users’ activities could still be revealed in the most
unlikely of ways. For example in one app that tracks sexual activity, the app makes specific requests to a certain
analytics service URL at the start and end of each session. In its communication, the app passes a unique ID for the
app instance and the app name itself as well as messages indicating start and stop of the tracked activity.
Based on this information, the third party who receives the data would be able to know the sexual habits of the
owner of the device, granted that the real identity of the device owner may not be associated with the ID. In this
case, the network exchanges were being made to request ads which will then be displayed to the user whenever they
are finished doing what they are doing. Because the requests for the ads are made in a deterministic way whenever
the activity is completed, it is possible for a third party to infer what the user has been doing. In addition, the app
also sends start and end messages to an analytics service provider which could enable a person with access to
the analytics data to determine the activities and performance of users of the app. Despite the makers of the app
promising that none of the users’ data is ever transmitted over the Internet, some of the users’ activity can still be
leaked through the network behavior of the app.
Aside from the scenario mention previously, there are also countless other scenarios where personal data could be
leaked unintentionally such as through human error or social engineering or just shoddy handling of data.

The quest for data
seems insatiable
and hi-tech
innovators are
constantly pushing
the envelope of
what and how
things can be
tracked.
WHERE IS SELF-TRACKING HEADING?

Where is self-tracking heading?
Despite the security challenges in self-tracking, public interest in it has mushroomed in the past few years
and there is no shortage of new startups and big players jumping into this space. One indicator of this interest
is in the amount of startup activity in this line of business. According to CB Insights, funding for quantified
self-related startups reached US$318 million in 2013, that’s up 165 percent from 2012. According to a report by
app analytics firm Flurry, the first six months of 2014 saw a 62 percent growth in the use of health and fitness
apps. Clearly this market segment is still in the rapid growth phase.
Another good indicator of where the market is heading can be had by observing what the major hi-tech giants
have been doing in this area. Many of them are making significant moves into this space. Facebook purchased
ProtoGeo, the maker of Moves in April 2014 for an undisclosed sum. Before the deal, Facebook already had one
of the most comprehensive and detailed databases of users, encompassing everything such as who their friends
and contacts are, what the user likes or dislikes, where they are going or have been, as well as their relationship
status and photos. By purchasing Moves, the social network adds detailed self tracking information to its
collection which gives it an unrivalled view into the lives of users.
Not to be left out of the picture, both Apple (HealthKit) and Google (Google Fit) have also announced major
forthcoming initiatives aimed at helping health and fitness category app builders tackle the challenges of
developing these types of apps. These initiatives will undoubtedly help to feed the growth in the health and
fitness category of the app market.
There will also be increasing crossover between wearable technology, the quantified self, and the Internet-of-
Things. As a result, the quantified self is increasingly morphing into something that is more akin to quantified
things, that is, the tracking of things owned by a person. Take for example Whistle – a tracking collar for your
dog or the quantified dairy cow using Lely T4C InHerd.
The quest for data seems insatiable and hi-tech innovators are constantly pushing the envelope of what and how
things can be tracked. Recent developments such as the MindRDR app for Google Glass sshow that we are not
too far away from being able to achieve the ultimate in life-logging and self-tracking — the reading and logging
of our own thoughts.
In recognition of the relentless trend towards collecting and using personal data by businesses, some individuals
are attempting to trigger a counter movement to regain the initiative in favor of the user or data subject. Take for
example the citizenme project founded by St John Deakins, which is created in an attempt to redress the balance
of power between those who are collecting and selling our data and ourselves, the data subjects.

Recommendations and mitigation
Both users and service vendors have a role to play in ensuring self-tracking security.
For users
The following steps could help users stay safe when using self-tracking apps:
• Use a screen lock or password to prevent unauthorized access to your device
• Do not reuse the same user name and password between different sites
• Use strong passwords
• Turn off Bluetooth when not required
• Be wary of sites and services asking for unnecessary or excessive information
• Be careful when using social sharing features
• Avoid sharing location details on social media
• Avoid apps and services that do not prominently display a privacy policy
• Read and understand the privacy policy
• Install app and OS updates when available
• Use a device based security solution
• Use full device encryption if available
For app developers and service providers
App and service providers should observe the following points to help provide a secure experience for users:
• Build security in from the start, not as an afterthought
• Always use secure protocols when transmitting data
• Ensure that the device is not directly or indirectly traceable
• Only collect data that is necessary to provide a service and nothing more
• Require strong passwords for user accounts
• Implement secure session management
• Follow best practices for password handling (only store salted hashes and not the real password)
• Follow secure coding practices
• Provide an easy to understand privacy policy and act within the stated policy
• Pen test system infrastructure to ensure security
• Ensure that backend systems are well protected from intrusion
• Make security testing a part of the product development process
• Ensure that staff are properly trained on how to handle sensitive information
• As a data controller, be sure to comply with relevant data protection laws

The self-tracking
craze is causing
an explosion of
personal data to
be generated,
transmitted, and
stored about
ourselves.
CONCLUSION

Conclusion
The self-tracking craze is causing an explosion of personal data to be generated, transmitted, and stored about
ourselves. Ultimately, the more data that we collect and store about ourselves, the more opportunity there is for
us to learn about ourselves, but it also opens up the opportunity for others to learn the same about us.
In this paper, we have examined some of the issues that can arise from the relentless rush to generate data
about ourselves. We have examined the types of self-tracking systems that are currently in use today and
how they generate and handle data. We have looked at the current state of security in the self-tracking space
and found it to be lacking in some key areas. For example many apps and services lacked privacy policies and
disturbingly, even basics such as the secure handling of user names and passwords are not done correctly by a
significant number of apps.
We also found that even devices that are not obviously traceable can still be tracked wirelessly due to
implementations that do not to use available privacy features.
With cloud-enabled systems, the user passes over much of the control and responsibility for safekeeping to the
cloud service provider who takes over custodianship of the data. The data held at the cloud service level has
a much wider scope of content and aggregates the data from all users of the app. The database could contain
data for millions of users and their activities. This places an onus on cloud service providers to ensure that they
implement the appropriate level of security and best practices to safeguard data integrity and privacy. Sadly,
as we have found in our research, the required level of care is not always taken, leaving users at risk. Service
providers should strive to ensure that security is at the core of the service from the device all the way to the
cloud. Security should be at the forefront rather than merely an afterthought.
So far, we have not seen large numbers of significant data breaches against operators in the health and fitness
app category but there has been some. As the sector continues to experience rapid growth, we can expect that
it will soon begin to register more prominently on cybercriminals’ radar and the question about possible data
breaches against major players in the health and fitness app sector is going to be about when, and not if, a
breach will occur.
Having scratched the surface of this burgeoning sector and glimpsed inside, we would conclude that there are
positive signs that some vendors are doing the right things, but far too many are not. Just how safe is your
quantified self? We think that it could be an awful lot safer than it currently is, so before you install the next new
self-tracking app on your smartphone or buy that new self-tracking device, pause for a moment and think before
you track.

Appendix
Models of self-tracking systems
The following are the most commonly used system models among self-tracking systems. The following table
summarizes the risk levels and various points at which data is collected, transferred, analyzed, and presented in
each of the common system models used.
Tracking device only
In this model, the data is collected by the tracking device and the
data stays on the device. The device fully handles processing and
displaying the data. There are not many modern self-tracking devices
that operate in this way. For example, old style pedometers work in
this way. They are typically not connected in any way and the data
stored within them is private and not shared with any third parties
who do not have physical access to the device. The limited display is
used to show all of the data and statistics captured.
Tracking device + smartphone app
In this model, the data is collected using a wearable tracking device. The data is then
transferred to an app that runs on the smartphone. The app is then responsible for
aggregating, analyzing, and embellishing the data, such as adding mapping, and
presenting of the information to the user. This model is used by some modern
self-tracking devices. Many of the wearable tracking devices do not have full
function information displays. Instead, they often just use a number of LEDs to
indicate essential status information only. The key to unlock the usefulness
of this type of system is the smartphone app, which processes and presents
the data back to the user.
Note: A smartphone in this context could also be a
desktop or laptop computer with Internet access.
Figure 12. Example of a basic wearable
tracking device, a pedometer
Figure 13. Example of a wearable device that transfers
collected data to a smartphone app
Table 2. Table of self-tracking system models
System model Privacy
risk
Data collection
point
Data storage Data transfer
mechanism
Data analysis Data
presentation
Tracking device Low Device Device None Device Device
Tracking device +
smartphone app
Medium Device Device
Smartphone
WiredWireless Smartphone Smartphone
Tracking device +
cloud
Medium Device Device
Cloud server
WiredWireless Cloud server Cloud server
Tracking device +
smartphone app +
cloud
High Device Device
Smartphone-
Cloud server
WiredWireless Smartphone-
Cloud server
Cloud server
Smartphone app Low Smartphone Smartphone None Smartphone Smartphone
Smartphone app +
cloud
High Smartphone Smartphone-
Cloud server
Wireless Smartphone-
Cloud server
Cloud server

Tracking device + cloud
In this model, the wearable tracking device collects data and
then the data is transferred directly to a cloud service which
is responsible for storing, processing, and presenting the data
back to the user. In some instances, the device may also be
able to perform limited processing and displaying of data to
the user but the real value is gained when the user logs onto
the cloud service to review the processed information.
Devices in this category can directly access the Internet to
transfer data.
Note: A wearable device in this context could also be a
stand-alone self-tracking device with direct Internet access.
Smartphone app only
In this model, an app is installed on the smartphone and the
app uses the various built-in sensors inside the smartphone
to track and monitor the activities of the user. Today, it is
common even for lower end phone models to have a range of
sensors such as accelerometer and GPS built in.
This system model does not use online cloud storage of the
data. Instead, all data is stored and processed locally. This is
a relatively uncommon way of doing things as most service
providers today are realizing the value of user data and are
aiming to capitalize on the data that their users are generating.
Using a relatively advanced device such as a smartphone to
collect, analyze, and display data is not a major disadvantage
for self-tracking and has advantages for privacy. However,
users may find it useful to view an analysis on a larger screen
or have the data backed up to the cloud in case there is a
need to upgrade to a new device or if the smartphone gets
misplaced.
Figure 15. Cardiio is a typical smartphone
only quantified self tracking app
Figure 14. Wi-Fi scales that can transmit data
readings directly to the cloud

Resources
Symantec Internet Security Threat Report 2014: Volume 19
http://www.symantec.com/content/en/us/enterprise/other_resources/b-istr_main_report_v19_21291018.en-us.
pdf
File Your Taxes Before the Fraudsters Do
http://krebsonsecurity.com/2014/02/file-your-taxes-before-the-fraudsters-do
Big Data Is My Copilot: Auto Insurers Push Devices That Track Driving Habits
http://business.time.com/2013/08/06/big-data-is-my-copilot-auto-insurers-push-devices-that-track-driving-
habits/
Social Networking Increases Burglary Risk
http://www.telegraph.co.uk/travel/travelnews/10171799/Social-networking-increases-burglary-risk.html
Wearable Technology – Market Assessment - IHS.com
http://www.ihs.com/pdfs/Wearable-Technology-sep-2013.pdf
The Tangled Web of Password Reuse
http://www.jbonneau.com/doc/DBCBW14-NDSS-tangled_web.pdf
Why Your Web Site’s Privacy Policy Matters More Than You Think
http://www.businessweek.com/smallbiz/running_small_business/archives/2009/08/why_web_site_pr.html
Anonymity, Privacy, and Security Online
http://www.pewinternet.org/files/old-media//Files/Reports/2013/PIP_AnonymityOnline_090513.pdf
Build Security In
https://buildsecurityin.us-cert.gov/
Common Weakness Enumeration
http://cwe.mitre.org/top25/
Handbook for Safeguarding Sensitive Personally Identifiable Information (DHS)
http://www.dhs.gov/sites/default/files/publications/privacy/Guidance/handbookforsafeguardingsensitivePII_
march_2012_webversion.pdf
Session Management Cheat Sheet
https://www.owasp.org/index.php/Session_Management_Cheat_Sheet
Here’s Why You May Never be Truly Anonymous in a Big Data World
http://www.nextgov.com/big-data/2014/07/heres-why-you-may-never-be-truly-anonymous-big-data-
world/88492/
Obligations of Data Controllers
http://ec.europa.eu/justice/data-protection/data-collection/obligations/index_en.htm

Bluetooth Company Identifiers
https://www.bluetooth.org/en-us/specification/assigned-numbers/company-identifiers
About Bluetooth Low Energy Technology
http://www.bluetooth.com/Pages/low-energy-tech-info.aspx
SIT Technical reports on The Security of cloud Storage Services
https://www.sit.fraunhofer.de/fileadmin/dokumente/studien_und_technical_reports/Cloud-Storage-Security_
a4.pdf
Data Dealers. Collecting, Collating, and Selling Personal Data Background Information and Research
http://datadealer.com/datadealer_backgrounds_research.pdf
Data Protection Laws of the World Handbook: Third Edition
http://www.dlapiper.com/en/us/insights/publications/2014/01/data-protection-laws-of-the-world-handbook/
Enhancing Cloud Security Using Data Anonymization
http://www.intel.com/content/dam/www/public/us/en/documents/best-practices/enhancing-cloud-security-
using-data-anonymization.pdf
Best Practices for Mobile Application Developers
http://www.futureofprivacy.org/best-practices-for-mobile-app-developers/

About Symantec
Symantec Corporation is an information protection expert
that helps people, businesses and governments seeking
the freedom to unlock the opportunities technology brings
- anytime, anywhere. Founded in April 1982, Symantec, a
Fortune 500 company, operating one of the largest global
data-intelligence networks, has provided leading security,
backup and availability solutions for where vital information
is stored, accessed and shared. The company’s more than
20,000 employees reside in more than 50 countries.
Ninety-nine percent of Fortune 500 companies are Symantec
customers. In fiscal 2014, it recorded revenues of $6.7 billion.
To learn more go to www.symantec.com or connect with
Symantec at: go.symantec.com/socialmedia.
For specific country offices and contact numbers, please visit our website.
Symantec World Headquarters
350 Ellis St.
Mountain View, CA 94043 USA
+1 (650) 527-8000
1 (800) 721-3934
www.symantec.com
Copyright © 2014 Symantec Corporation. All
rights reserved. Symantec, the Symantec Logo,
and the Checkmark Logo are trademarks or
registered trademarks of Symantec Corporation
or its affiliates in the U.S. and other countries.
Other names may be trademarks of their
respective owners.
Any technical information that is made available by Symantec Corporation is the copyrighted work of Symantec Corporation and is owned by Symantec
Corporation.
NO WARRANTY . The technical information is being delivered to you as is and Symantec Corporation makes no warranty as to its accuracy or use. Any use
of the technical documentation or the information contained herein is at the risk of the user. Documentation may include technical or other inaccuracies or
typographical errors. Symantec reserves the right to make changes without prior notice.
Authors
Mario Ballano Barcena,
Candid Wueest,
Hon Lau

WHITE PAPER: How safe is your quantified self? from the Symantec Security Response Team

More Related Content

Viewers also liked (8)

Similar to WHITE PAPER: How safe is your quantified self? from the Symantec Security Response Team (20)

More from Symantec (20)

Recently uploaded (20)

WHITE PAPER: How safe is your quantified self? from the Symantec Security Response Team