IN THE FEDERAL TRADE COMMISSION OF THE UNITED STATES
IN RE
HE MOBILE WIRELESS WEB, DATA SERVICES AND BEYOND:
EMERGING TECHNOLOGY AND CONSUMER ISSUES
A Public Workshop
RESPONSE STATEMENT FOR DAY II PANEL:
Building Privacy and Security Solutions into the Technological Architecture
Gregory A. Miller
Chief Privacy Officer &
Vice President Corporate Development
MEconomy, Inc
1207 Indiana Ave Suite 1800
San Francisco, CA 94107
gam@meconomy.com
11.December.2000
I. Introduction
Mr. Chairman, Commissioners, Commission, and the Bureau of Consumer Protection: My name
is Gregory Miller. I am an Internet business strategist whose 20-year career blends
education and experience in technology, business, and law. My professional pursuits have
tracked the evolution of the Internet from the days of the Arpanet to the engine of
today's digital economy. My accomplishments are divided between software engineering and
marketing/business development, with experience as a technology lawyer and an Internet
public policy specialist. I have previously served on an FTC special committee for
consumer online privacy and act in an advisory capacity to US Senators on matters of
Internet policy. I also serve on the public policy committee of the Critical
Infrastructure Assurance Partnership. In the private sector, I currently am the Chief
Privacy Officer and Vice President of Corporate Development for MEconomy Inc. In this
capacity, I am charged with ensuring MEconomy's product strategies fully support best
practices and current laws governing consumer privacy. As a member of the executive
management team I participate in setting business strategy and developing strategic
alliances and relationships. This includes serving as MEconomy's thought-leader on
consumer Internet privacy and security and liaison with government. My complete curricula
vitae are incorporated by reference herein.
MEconomy is an emerging technology upstart company whose business is rooted in
providing the technology infrastructure of an assured privacy layer
("APL") for the Internet. This APL will work across all delivery channels, wired
and wireless. Within the context of the wireless medium - the focus of the FTC's research
inquiry herein - MEconomy will offer an innovative technology for wireless gateway
providers and intermediaries, among other types of service providers in the digital
economy. MEconomy's system will offer an Infomediary that creates, maintains, and protects
device, location, user, and service profiles. This Infomediary will empower the consumer
to directly control their own data while allowing them to authorize the use of their data
- giving permission to destination sites and other service providers - for specific and
pre-identified purposes on an "opt-in" basis. MEconomy currently has portions of
its software infrastructure in "alpha" and "beta" testing phases, and
expects to launch a production-ready commercial-grade solution in the second quarter 2002.
On behalf of MEconomy and myself, I wish to extend my gratitude for being selected for
this FTC undertaking and the opportunity to participate. The Federal Trade Commission has
provided a set of issues and questions pertaining specifically to the building of privacy
and security solutions into the technological architecture. On behalf of, and representing
the business interests of MEconomy, Inc, as respondent and panelist member I respond to
those questions for the record herein, preceded by a brief wireless industry background
statement to set the background for my answers.
II. Wireless Industry Background Statement
The convergence of voice and data with delivery across a wireless medium should affect
many sectors of the economy. I believe that a major software infrastructure will be
required in the telecom sector to manage, route, and filter this traffic as it converges.
Entirely new software applications and services will be developed to empower wireless
users. Of course, within this infrastructure there must be new layers to manage privacy
and security. Therefore, I offer the following observations as a backdrop for the answers
to the questions posed by the Federal Trade Commission about privacy and security
solutions into the technological architecture.
- Infrastructure - The first wave of wireless data services will be the
infrastructure itself. With any emerging platform, the first wave of investment is
typically in the underlying infrastructure that is required to deploy services. We are
hopeful that a component of this will be privacy and security mechanisms that foster trust
within eCommerce delivered across the wireless medium. I believe the wireless
infrastructure build out will continue for the next 12 to 18 months.
- Corporate Focus - Over the next 12 months the focus should be on the
corporate market rather than the mass consumer market. In the next 12 months the market
will likely be driven by early adopters, and I believe that the mass consumer market will
not significantly contribute to adoption for at least a year due to the lack of compelling
content and applications. I could add the lack of privacy and security to that
observation; however, consumers have little reason to give significant thought (yet) to
the degree of privacy and security that lies in their mobile device. Once digital wireless
devices entered the market, the general scare of analog wireless devices being
"jacked" by telecom criminals standing on overpasses of major thoroughfares
using technology available over the counter, the concerns of trust in the system integrity
largely disappeared. As will be disclosed in my answers infra, this lack of
concern (or awareness) likely will change as fresh light is shed on the blatant
vulnerabilities of the current digital infrastructure integrity. However, for the next
year, the focus will largely be on services provided in the corporate setting - I believe
largely because of the market's recent scare about the adoption risk inherent in
subscription-based business models. Market research and analyst reports suggest that
carriers are developing wireless services to offer to their enterprise customers over the
next six months in an attempt to drive revenues prior to consumer adoption. It is during
this stage that efforts should also be plied in the lab to bring about orders of magnitude
improvement in system integrity for consumer privacy and security when that adoption
begins.
- Growth Challenges - Over the next 12 to 18 months wireless data
services adoption rates, usage patterns and revenue models will vary by geographical
region. This is due in part to the frustrating state of standards in the United States
compared to elsewhere on the globe. And I do not believe that Japanese NTT DoCoMo's
wireless data penetration, which was initially driven by entertainment applications,
should be used as a proxy in the United States.
- Market Adoption - Wireless data services represent a large market
opportunity, but there will be roadway hazards along the journey to profitability and
wide-scale adoption. Worldwide cell phone subscribers are forecast to increase to more
than 1.6 billion by 2004 up from 475 million in 1999, representing a 28% CAGR, and
worldwide wireless data subscriptions will increase to more than 400 million by the end of
2004 up from approximately 15 million at the end of this year, 2000, representing a 127%
CAGR.(1) However, I believe that this subscriber
penetration could be adversely affected by technological constraints, lack of compelling
applications, and most importantly, renewed fears of lack of device privacy and security,
or system integrity.
- Application Trends - There will be a dramatic difference between
wireless and desktop Internet applications. Wireless data services must be tailored
specifically to the wireless device, and different devices will serve different markets.
Location-specific services will likely permeate all wireless data services, similar to the
ubiquity of search engines for the web on the desktop. Along these lines I believe five
trends will emerge: 1) the convergence of voice and data; 2) growth of personal
information management applications; 3) improved synchronization technologies; 4)
optimized delivery platforms; and 5) an explosion of location-based services.
- Success Attributes - There are four attributes for success of wireless
data services: ubiquity, integration, utility, and simplicity. And much of the
functionality of data services will be made possible by voice-enabled technology (e.g.,
TellMe Networks - http://www.tellme.com) because
today's 475 million wireless devices are optimized for voice communications, not data.
- Technology Constraints - The wireless Internet experience will be
constrained over the next 18 to 24 months. The desktop is a powerful platform for
information gathering, research, and commerce. Building privacy and security
infrastructure around that is comparatively easy. However, the wireless platform is
constrained by several factors including: lack of processing capabilities, limited
bandwidth, network latency, limited user interface real estate, limited user input
capability, the slow transition from switched to packet data transmission, closed carrier
platforms and multiple standards, and limited applications.
- Customer "Ownership" - Some bristle at the use of the term
"ownership" in conjunction with the word "customer," however, the fact
remains that existing desktop brands and new entrants will compete with carriers for
customer loyalty. All are likely to have to share revenues, as the platform becomes
mainstream and transaction revenues begin to channel through the wireless medium. I do not
believe the leading desktop brands such as Yahoo, Excite@Home,
or NBCi will become dominant wireless Internet brands unless the content and services they
offer quickly scale and tailor to the constraints of the wireless platform outlined supra.
III. The Specific FTC Questions & Answers
- Are there basic principles of technology and design
that have emerged in the Internet space that can be applied in the wireless context to
increase privacy and security?
For example, consider the basic principles of minimizing the degree to which data is
personally identifiable; keeping information in the custody of the consumer to the extent
possible; making data flows transparent/visible to consumers; encryption of all data and
meta-data; minimizing data collection and retention; and decentralization of databases.
From a practical and policy standpoint the principles of technology and design that
have emerged to manage or increase security and privacy are not always in parity with the
same ones that have been adopted by the mainstream, and this is discussed infra.
Further, there is not, in our opinion, a discernable amount of empirical evidence that the
principles cited in the example above or some principles we add below have "increased
privacy and security". For example, giving consumers "complete control
of their information" is a noble concept, but we do not believe there is
sufficient evidence to conclude that this directly increases any privacy or security.
With that in mind, here are five principles, or perhaps goals we believe are worth
consideration for privacy and security in the wireless infrastructure:
- The user's direct access device (e.g., wireless appliance, etc.) should be the initial
source of encryption, and it should be performed end-to-end, only with trusted
intermediaries.
- The end device should be an open platform so users can load (or unload) their own
privacy/security technologies.
- All technologies that directly and initially touch consumer's data (e.g., at the point
of collection) should be available for unfettered public review (i.e., we support open
source initiatives).
- Consumers must have complete control over their information (e.g., fair information
reporting principles of notice, consent(2), access, etc.).
- Any data collected for a transaction should be de-coupled from personally identifiable
data, and only used for that transaction.
We believe that by and large, consumers are not completely aware of the potential
problems with privacy in a digital economy (e.g., identity theft, discrimination, crime,
etc.) until a disaster strikes home.(3) Likewise, consumers
want a transparent experience that they can trust and rely on. For instance, by and large,
the lock icon that has become a fixture of web browsers to indicate that data is
transferred securely has become a visual cue that requires no further intervention by the
user (save a mouse click to acknowledge an alert panel). We believe a similar approach
needs to be taken with regard to privacy - creation of an "assured privacy layer,"
if you will. However, this discussion with regard to privacy in the context of the World
Wide Web is significantly impacted by the constraints of the wireless medium. So we need
to consider this question in light of the technical landscape of the mobile or wireless
channel.
From a technical standpoint, good security requires the right tool for the right job.
In other words, we need to use the proper application of the most effective protocols and
processes. The four corners of consideration include:
- User interfaces
- Processing power
- Memory
- Bandwidth
Arguably hardware innovations have not arrived faster due more to business challenges
than technology limitations. Likewise, the state of wireless security is less conditioned
on hardware limitations and more on externalities explained infra. At the same
time, we find that wireless application developers are asked to implement security
capabilities suitable for desktop computing but not ready for wireless devices due
primarily to these four architecture elements supra. As hardware improves, we
believe security mechanisms may likewise improve, assuming we can circumvent the business
and political challenges.
The most promising cryptography scheme for wireless clients appears to be ECC
(Elliptical Curve Cryptography), a variant of public key cryptography algorithms. Strictly
speaking, it is still under technical review, but is considered promising and industry
consortia like the SECG are actively promoting implementation of ECDSA in mass-market
products.
ECC provides security (and ECDSA provides authentication) comparable to that of RSA or
Diffie-Hellman (of which the Elgamal variation is most widely used for encryption). The
security of the hardware is still an issue, but that is continuing a problem with mobile
systems that rely on discrete hardware to store secrets.(4)
Let's consider each point in the example provided by the Commission.
- Minimizing Identifiable Data
In terms of "minimizing the degree to which data is personally identifiable,"
Pseudonymization (e.g. AMEX' "Blue" card) is becoming "acceptable" in
the business world and could become commonplace in 2-4 years. This is because technically
speaking; pseudonymity is built from security primitives with "provable"
characteristics.(5) Pseudonymized networking protocols
(e.g. Zero Knowledge Systems re-mailer) in some winning format (not necessarily Zero
Knowledge Systems' solution), will be widely used. Law enforcement will always seek
over-access, creating healthy tension that drives the technology curve. For political and
economic reasons, we believe that the EU will be a primary catalyst for this tension.
- Keeping Information in the Custody of the Consumer
Custody is important, but we believe the custody issue goes to more to the keys than
the data itself. We believe one reason the security community rebuffs policy proposals for
any form of key escrow is the concept that the root of privacy lies in the custodianship
of the keys. In other words, keeping one's keys anywhere but with the individuals
themselves proffers compromise on its face.
- Making Data Flows Transparent/Visible to Consumers
We believe it is more important to make the security mechanisms transparent than it is
to making the data flows visible. And we believe this is a driver for an increasing trend
to place security software in an open source environment, subject to peer-review and full
disclosure. The real issue, we believe is making certain data transport mechanisms are
actually as secure as purported (rather than - for example - making certain a clickstream
is opaque). To quote a former U.S. President, "Trust and verify."
- Encryption of all Data and Meta-data
Encrypting data is a good measure, however, the challenging part, and the real focus
ought to be verifying that the authorized individuals and only those authorized have the
necessary keys. Assuming, however, that key management is properly maintained, encryption
of data should be a business and/or utility proposition, predicated on the corresponding
issues of ease of access, overhead, usability, performance, cost, etc. Simply encrypting
all data will not solve the challenges of data protection, and in particular, the real
issue is key management and security.
- Minimizing Data Collection and Retention
The Internet has provided a means for new levels of surveillance, unprecedented
capabilities in direct marketing, and innovative approaches to customer care. And there
are charges that what has evolved is more than a commercial Internet, but a global
surveillance system of national and international security import. As a result, regardless
of which purpose one subscribes to, the most difficult adversary may lie beyond unwanted
direct marketing. It may lie within the constant tension that strains the relationship
between individual privacy and the government's charter to protect its nation. These
practices are increasing (e.g., see the Walsh Report).
We believe that data collection should be minimized to purposes that must be disclosed.
However, philosophically, we believe it is equally important to be wary of technologies
and infrastructure that can be (and in some cases are) deployed for surveillance purpose -
particularly where there is no clear and present need under well settled federal
wiretapping and eavesdropping statutes. And regardless of whether technology is being
deployed for such purposes by law enforcement or government agencies, we must be mindful
of the illegal deployment of such technologies.
Therefore, when considering data collection and retention - regardless of the medium,
wired or wireless - it is perhaps first and foremost to consider the regulation of data
collection and retention. Currently, there is no regulatory or statutory scheme to
minimize data re-marketing. We do not mean to suggest that regulation is the answer to all
of the privacy and security issues or questions, however, we're also mindful that in some
instances an appropriate mix of industry initiative and government support is in order.
- Decentralization of Databases
Centralization is a difficult proposition to resist, but peer-to-peer technology may be
the natural antidote. Distributing data with strong crypto protection may afford a
competent, feasible, and consumer comforting solution. Consider how gateway providers and
intermediaries might maintain the aggregate data they want and need to render their
services, but allow the actual profiles to remain directly in the control of their
rightful owners - the consumer. On the other hand, in some settings, this may prove less
than practical. Accordingly, we believe it comes down to having a strong trust
relationship. If a trust relationship exists, then the aggregation of data into
centralized databases may not be a pressing matter.
- How are companies wiring in privacy and security today
in the wireless medium? Are the protections currently employed adequate?
Simply put, we believe companies are racing to incorporate security advances that we
believe are not being completely analyzed for security integrity. And as a result, the
current employed protections are inadequate.
The supporting authority for this answer is highly technical. An overview is
incorporated infra for the Federal Trade Commission's reference. Verification of
the assertions below may be found in the literature or through review from security
experts. Accordingly, the following details are based on descriptions provided by security
experts whose general knowledge exceeds that of the author, but is readily available in
the technology literature.
Generally, we believe the protections currently employed attempt to strike a balance
between "breakability" for government intelligence agencies and sufficient
security to thwart casual intercepts, and are reasonably justified by limitations in the
technology (the four corners of consideration outlined in Question 1, supra).
For example, some academic cryptanalysts have argued that the ten+ zeroed-out bits in
the A/2 and A/5 ciphers specified by the GSM Consortium is a classic example of this,
although the wireless telephony industry denies this report. Add to this that the wireless
security landscape has arguably changed dramatically in the past decade. Powerful
cryptanalytic capability is so widely available in the intelligence world that it has
naturally percolated down as publicly available computing techniques and inexpensive but
powerful hardware. In fact, the types of intrusions that can be performed by relative
amateurs have admittedly surprised law enforcement agencies from time to time, but they
shouldn't.
One initiative containing these trapdoors includes WTLS, a wireless implementation of
the Transport Layer Security (currently with known security flaws that competent security
experts can explain). In the authentication area there has been good work using sufficient
key lengths, such as the use of ECDSA (Elliptical Curve Digital Signature Algorithm) in
wireless protocols for authentication. The key lengths are not foolproof of course,
because any good authentication algorithm can also be used to encrypt (e.g. the
Elgamal encryption variation on the Diffie-Hellman key exchange protocol).
TLS and WTLS bear a close resemblance to each other because the WAP Forum incorporated
large sections of the TLS Working Group's protocol work directly into the WAP
specification. This was a useful step to facilitate rapid development of the wireless
medium and push the specification forward. However, the WAP Forum arguably also introduced
several major cryptographic vulnerabilities to TLS. These compromises may be due to
standard design constraints on wireless systems including:
- Mobile unit processors lacking power and memory capacity
- Bandwidth limitations
- Export restrictions on cryptography (as a market penetration factor)
- Common long round-trip latencies
- Lack of support for both connection-oriented and datagram transport layer protocols
Diving deeper into the technical details, the results are weaknesses in current
wireless protocols using implementations of the WTLS security schema (e.g., WAP).
Specifically, security experts claim that adding support for datagrams, optimizing the
packet sizes for poor quality wireless environments (e.g., limited bandwidth) and
specifying faster algorithms all introduced weaknesses to several well-documented attacks,
including the following:
- WTLS reduces the key space by a factor of 32 from 40-bit DES by using 35-bit DES
encryption to make it weaker and thus more exportable. 56-bit DES has 8 bytes of keying
material, and therefore, we believe WTLS is unable to meet its own requirements for
"best possible security."
- Another more serious allegation regards the "record_type" field. This data
field is sent in the clear, so key changes could be eavesdropped and substituted. One
example is the "ChangeCipherSpec" field. An attacker could turn encryption off
by switching an unsuspecting owner's WTLS enabled device to a null cipher without the
owner's knowledge. If true, this is a serious security flaw.
- Another allegation is that alert messages are sent in the clear and unauthenticated.
Since they're given a sequence number, an attacker can - without detection - selectively
replace encrypted datagrams with unauthenticated plain-text alerts by assigning the same
sequence number. The result is what is referred to as a "truncation attack"
where arbitrary packets can be dropped from the data stream selectively, or more
sinisterly, replaced with others.
- Block ciphers are used in CBC mode (cipher block chaining), but WTLS uses a 40-bit XOR
message authentication code (MAC), which "pads" message blocks with zeroes and
then divides it into 5-byte "blocks" to be "XOR-ed" (a mathematical
recombination process). Regardless of these short key lengths, this use of a weak XOR MAC
in the stream (block) cipher in this manner provides no integrity protection.
- The block ciphers in WTLS are highly vulnerable to "brute force" attacks. To
find the correct key, an attacker runs a trial decryption on the final block of each
packet. Given the inexpensive processing power of laptops and other small devices commonly
available today, brute force attacks no longer require a roomful of compute power.
- WTLS specifies Diffie-Hellman key agreement computations using 512-bit and 768-bit
primes with generators, but it does not specify the group order generator of the
multiplicative subgroup. This suggests one cannot check if the given public value belongs
to the correct subgroup, which means an attacker has a backdoor if they want to mount a
"factorization" attack. This RSA encryption is vulnerable to chosen
"cipher-text" attacks. Security experts allege that cryptanalysis can be
successful in approximately 220 rounds - insignificant by today's commonly
available compute power - because the RSA signatures and encryption conform to PKCS#1
v1.5. That standard provides an oracle that reveals if the correct padding is present on
any given packet. And some WTLS implementations have error messages like
"bad_certificate" and "decode_error" that can easily provide an
attacker with precisely such an oracle.
In summary, chosen plain text (data recovery) attacks, datagram truncation attacks,
message forgery attacks and "exportable" key space searching shortcuts, are all
serious privacy problems buried in the technical details of today's digital wireless
device implementations using WTLS v1.0. WTLS v1.1 has not remedied any of these
allegations. The new specification implicitly suggests solutions, but fails to implement
them (e.g., SHA as an algorithm for handshake protocol).
Unfortunately, a similar story exists with the GSM security specification.(6) It is alleged that 150+ million GSM users worldwide are
subject to eavesdropping, spoofing, cloning and hijacking and every single major GSM
provider have been less than forthright about the sufficiency of security, marketing PCS
devices using GSM technology as "secure" (e.g., Pacific Bell, a subsidiary of
SBC). Arguably, some experts may counter that the security breach potentials outlined
herein and alleged in the literature are so deminimis as to (as a matter of business
decision and strategy) not be worthy of resolution at this time. However, there are
security experts who claim to have seen demonstrations of commonly available desktop
hardware and software technology that cryptanalyze and clone a PacBell PCS/GSM phone
within 3 hours.
To take this one step further, we believe this form of identity theft is simply waiting
to happen. Consider that the cloning approach mentioned above can also be deployed as a
masquerading "base station," so an attacker could sit in a convenient public
location, and abscond with the A/5 keys of every PCS phone user who passed by talking on
their PCS device and within a few hours, have cloned their phone. Security experts allege
this to be true in the face of PacBell staunch denial, but regardless of what is the
truth, we believe there is enough in question among reasonable persons to call for a
deeper audit of current security architectures before any policy or regulation is
fashioned.
- Are there additional existing technologies that could
build privacy and security protections into the architecture of wireless devices and
services (For example, XNS or others)?
We first note in passing that XNS looks most promising. In general, however, we believe
wireless privacy and security is not a technical problem, but more of a political problem.
For example, the IETF (Internet Engineering Task Force) IPsec working group, the ANSI
(American National Standards Institute) X9.F1 working group and the IEEE (Institute of
Electrical and Electronic Engineers) P1363 working group have adopted Diffie-Hellman
(X9.42), TripleDES Modes (X9.52), and ECDSA (X9.62, an elliptic curve version of NIST's
Digital Signature Standard). Meanwhile, the wireless industry is still waiting for the WAP
Forum to specify strong security mechanisms that overcome the current alleged
deficiencies. We believe that if the current protocols had all the alleged weaknesses
addressed, in conformance with the repeated recommendations of the academic and technical
security communities, 95% of the most serious privacy vulnerabilities would be removed.
On the other hand, resolving these vulnerabilities would make law enforcement and other
government agencies' work more difficult - the political counterbalance. As a result, we
query whether millions of wireless device owners use weak devices because there remains a
threat of bad actors relying on the wireless medium to perpetrate their misdeeds.
Unfortunately, if so, then this situation twists into an irony because the ability to
compromise the existing security infrastructure is now within the reach of amateurs.
- What kinds of technologies, standards or models may
emerge to safeguard privacy and security in the wireless area? How far away are they?
(For example, could a system be designed where consumers' protections and permissions
travel with their data to protect against misuse or unauthorized disclosure?)
As we've stated, these technologies, standards, and models to a large extent exist, but
political process may be inhibiting their adoption. In terms of models or standards, we
are proponents of creating a privacy infrastructure that:
- Relies on peer-to-peer networks, but supports aggregation of authorized data
- Empowers a consumer with an opt-in permission capability
- Creates an assured privacy layer using proxy technology and a sound PKI strategy
- Given the limitations that exist in terms of bandwidth
and computing power in the wireless area, is it possible to have strong security and
privacy without compromising ease of use and speed? What kinds of trade-offs will
consumers need to make between convenience and protection?
We believe the limitations cited are not a good excuse for the choices made in
specifications like WTLS. Lengthening keys by a few bytes would not seriously impact
performance, especially in the newer phones and PDAs. Properly implementing better
algorithms could, in some cases, make the protocols more efficient. Choosing better and
longer initial vectors without zero padding and transmitting them encrypted would not be
as large an overhead problem as suggested (it's the same amount of data in most cases).
Using ECDSA would add authentication and prevent eavesdropping and truncation attacks. We
believe the newer processors can easily handle these things.
It remains to be seen if the next version of the specifications incorporates better
algorithms like SHA-1 and ECDSA, longer key lengths, and protocol improvements such as
encrypting alert messages. And it remains to be seen when and how the manufacturers will
incorporate faster CPUs. There are business case arguments that consumers will not pay
extra for these security features, but with the rise of public concern over privacy and
security this may change as well. Having said this we still believe it is possible today
to have strong security and privacy without compromising speed and ease of use. Compute
power/price ratio is continually improving. If existing technologies are re-evaluated to
employ solutions to some of the pressing problems cited supra, then the industry
can achieve better levels of privacy and security. However, there are consumer tradeoffs.
The two most significant are:
- Price/Performance. Security technologies require more on-board hardware
capability, which can raise costs. Accordingly, the "security-enabled" wireless
device will generally cost more than the unprotected device. And until there is either
regulation or a terrible catastrophe (e.g., identity theft that gains national or
congressional attention), consumers will not spend the required time, energy, or thought
processes to make the "right" decision, and the unprotected devices will win
out.
- Style/Performance. Small is in, and getting smaller by the month. With
a reduced footprint for hardware, security feature/functions may be the first to
trade-out. In the rush to get the smallest, flashiest device, consumers will have to make
a personal information tradeoff. Unfortunately, similar to home security systems, until a
loss is sustained, consumers won't spend the time, money, or thought processes evaluating
this tradeoff. The result is that product manufacturers will not be driven to incorporate
higher grades of privacy or security until a business case can be established, and the
business case won't be established until there is market demand. Market demand, as noted
will require consumers understanding or experiencing a clear and present need.
- Are there limits to the amount of privacy and security
protection that can be achieved technologically? If so, what are the other elements needed
to ensure consumers' privacy?
For example, what might be the role of self-regulation, government regulation, market
forces, civil law/liability, or employee training?
Yes, and we believe there are four elements that must be addressed to better assure
consumer privacy.
- Key Management - People will forget their keys or give them away.
- Education - Unless consumers understand the risk to their personal well
being in an unsecured transaction mechanism such as the ones being deployed today in the
wireless space, price and style will always win out.
- Regulation - While we believe that market forces will be sufficient
industry motivators, we also believe there is an important role for reasonable and
appropriate government partnership and initiative through prudent legislative and
regulatory guidelines.
- Incentives - MEconomy is founded on the belief that unless the
incentives of the advertisers, the marketers, and the consumers are aligned, these privacy
and security concerns will continue. Consumers have to be motivated to care about the
control over their personal information. Advertisers have to be motivated to care about
not compromising consumers' personal information. Marketers have to be motivated to not
gather and misuse information. A system like MEconomy's assured privacy layer rewards
consumers for authorizing specific uses of their data - giving permission to destinations
and other Web services to leverage their data on a sliding scale of rewards, incentives,
and other benefits. At the same time, this same system must (and will) provide a layer of
assured privacy.
1. See Thomas Weisel Partners Wireless Infrastructure Analyst
White Paper, Matt Finick, 28.August.2000
2. The concepts of "choice" and "consent"
have been interchanged in recent literature. The respondents recognize that
"consent" suggests an "opt-in" strategy, whereas "choice"
could be either. We're further appreciative of the fact that mobile commerce may become
heavily weighted with financial services offerings. And under provisions of Graham Leach
et al legislation the strategy is clearly "opt-out." We recognize this may
create an impasse - one the FTC should plan on addressing sooner than later because the
public policy momentum appears to be building within the wireless business sector to
support "opt-in" strategies.
3. Although unproven in the marketplace as yet, we further
believe that an adopted Infomediary acting as the consumer protectorate would go a long
way toward increase privacy and security.
4. Due to power/timing attacks on the key schedules.
5. By "provable" I mean to suggest that technical
proof notwithstanding, eventually, policy makers realize that mathematics works.
6. For details on these vulnerabilities see http://www.scard.com regarding GSM vulnerabilities
particularly concerning the US version of GSM. |