Are Clouds making Our Research Irrelevant and Who Is at Fault?
(Position Paper)
Yvo Desmedt
a
Department of Computer Science, The University of Texas at Dallas, 800 W. Campbell Road, Richardson, U.S.A.
Keywords:
Information Security, Cryptography, Clouds, Research, Deployment.
Abstract:
Until recently, the user of a computer system was able to (at least to some degree) help decide security poli-
cies, such as which access and information flow control to use, which cryptographic algorithms to choose,
how to secure databases in use, etc. Due to these choices, researchers were able to have an impact on what
was deployed.
In today’s world, the Chief Information Officer (CIO) outsources online communication (replacing landlines),
databases, e-mail, storage, voting, WWW, etc., to clouds. These do not use open source and do not disclose
their design. So, the security is left to the designer and the user is completely left in the dark. Since most
programmers never took a course in information security, we should assume the worst.
In our paper we justify several positions: (i) we make the claim that clouds have lowered our information
security; (ii) we wonder whether CIOs compare competing clouds on their security properties and ask inde-
pendent experts for their advice; (iii) one finds that self-acclaimed experts often lack basic knowledge; (iv) that
research is becoming irrelevant. We also wonder who is at fault for these problems and how we can address
them.
1 INTRODUCTION
Some well known attendees of the Crypto conference
announced they would no longer attend the confer-
ence because it had become irrelevant to real world
problems. Indeed, despite having been co-editor of
the 1982 Crypto proceedings (Chaum et al., 1983),
Alan Sherman (PhD, MIT) was one of the first to
make such a statement. Another example is Paul van
Oorschot, who became involved with USENIX. Fi-
nally, IACR (who organizes Crypto) created “Real
World Crypto” to address such concerns.
When we compare computer security with cryp-
tography, we can hardly say that the area is dominated
by theoreticians. In this paper, we will argue that a
lot of the research on computer security is irrelevant
to the real world today. We argue that the computer
world in which we live has dramatically changed. 15
years ago any mid-size to large organization (whether
business or non-profit) would run their own servers
and data would be stored locally. Today, we usually
find a cacophony of cloud servers. We will put for-
ward the position that: “the computer environment
has changed, but many researchers have not adapted
a
https://orcid.org/0000-0002-6679-7484
their topics, making the research irrelevant.
Before we justify our positions, we first wonder
how we came to a world in which cloud servers have
taken over the role of local servers. For this, we start
in Section 2 by considering the history from Win-
dows 95
1
and dial-up internet
2
on. We also explain
the dramatic impact these had on modern computer
systems. In Section 3 we give concrete examples
how during the last 15 years cloud servers were se-
lected. In Section 4 we try to talk about cloud se-
curity. Indeed, since many clouds in use today are
closed source, we wonder what we actually can say
about their security. In particular we focus on whether
information security concerns, such as privacy, re-
ceived the attention they should have, when moving to
a cloud based world. Before we question whether aca-
demics should study the security of particular clouds
(see Section 7.1), we state our positions in Section 5.
Finally we wonder In Section 7 who is at fault.
1
See: https://en.wikipedia.org/wiki/Windows\ 95
2
See: https://en.wikipedia.org/wiki/Dial-up\ Internet\
access
436
Desmedt, Y.
Are Clouds making Our Research Irrelevant and Who Is at Fault? (Position Paper).
DOI: 10.5220/0011328800003283
In Proceedings of the 19th International Conference on Security and Cryptography (SECRYPT 2022), pages 436-442
ISBN: 978-989-758-590-6; ISSN: 2184-7711
Copyright
c
2022 by SCITEPRESS Science and Technology Publications, Lda. All rights reserved
2 THE RECENT HISTORY OF
COMPUTERS REVISITED
3
In the 1990’s very few individuals had a good internet
connection. Users needed to dial (phone) an inter-
net provider to connect to the internet, blocking the
phone! So, PCs at home were offline most of the time
and often switched off. As a consequence, the dial-
up internet providers hosted e-mail accounts and web
pages. So, there was no need to have user friendly
software for installing e-mail and web servers.
With the appearance of cable and DSL modems,
users being on the internet 24 hours, the history of
computers should have changed dramatically, but it
did not. We now look at the post dial-up internet
world and describe what could have happened in a
parallel universe.
First 24/7 internet connection could have allowed
for users to have e-mail and web servers at home, as
we now explain. Although PCs are turned off after
use (PCs are noisy), home user’s modems are on 24
hours! We now argue they could have hosted these
servers. Modern “modems” are more than strictly
modems. They also contain a firewall, router, etc.
Moreover, the interface with the user is a web in-
terface! So, a different design could have consisted
of using the modem for a web server providing a
basic web page, and internet providers could have
sold/rent “sophisticated modems” that enable more
advanced web pages. Moreover, for e-mail, the in-
ternet provider could have acted as a backup for in-
coming mail, which would had been forwarded when
the home user’s PC comes online, in a similar way as
POP allows.
Unfortunately, these who designed modems,
never provided the aforementioned service, and so
when DSL, coax, fiber, etc., allowed for 24 hours ser-
vice, users regarded it as normal that they needed an
external e-mail address and using external tools, such
as what today is called social networks, to dissemi-
nate information, which could have been done by us-
ing local web servers. So, today’s social networks are
centralized, while they could have been distributed,
avoiding a whole range of problems, such as the cen-
soring of Facebook, LinkedIn, Twitter
4
, etc.
Another potential use of the home modem could
have been to set up a VPN. This could have been used
3
This section is based on the author’s seminar lecture
Is The Rise of Cloud Storage, Cloud Computing and So-
cial Networks a Consequence of a Failed OS (Operating
System) Design? at Microsoft Research, Cambridge, UK,
on November 27, 2013.
4
The writing of this paper started before Elon Musk
considered buying Twitter.
to help reroute data transmission. For example, when
two persons travel to Japan, one a UK resident and
the other a US one, their respectively home modems
“know” they are in Japan. Indeed, the home modem
would had been a basic home server. So, data that
needs to be communicated between these two people
does not need to travel via the US or the UK, leading
to a distributed communication, instead of the cen-
tralized ones we have today. This approach seems far
fetched, but mobile phone companies have been doing
exactly this for decades!
The 24 hours internet could have had some other
impacts, which OS (Operating Systems) designers did
not realize, as we now explain. Indeed, they failed to
update the concept of “user. The definition of user
should be: Anybody who uses (or should use) re-
sources on your computer”. That implies that when
hosting a web server, the whole world is a potential
user! Note that when being connected 24/7, from a se-
curity viewpoint, anybody should also be called a (po-
tential) user! When we use this corrected definition of
user, sharing data with a particular person (anywhere
in the world) should have been made extremely user-
friendly. However, since the concept of user was not
adapted, this never happened, and so today clouds are
being used to share data.
The anti-cloud world we described may seem un-
realistic. Indeed, search engines contain so much data
it can not fit on a modern PC/laptop. However, since
inexpensive disks today can contain 2-3 Terra Byte
(TB), a lot of information could be stored locally.
Such disks could have low resolution maps, a lim-
ited Wikipedia (e.g., without pictures) and for pro-
fessional users, data they might need. For example
the proceedings of a conference today are typically
25Mbytes. So, a 3 TB disk could contain 120,000 of
these proceedings, which is more than what is needed!
The companies selling disks have not understood they
could preload disks with useful data!
Anyway, we do not live in this parallel universe.
Modem designers, OS designers, etc., never under-
stood they could have provided an alternative to our
cloud-centered world. To better understand how we
came to this world, we illustrate in the next section
how some decisions were made.
3 HOW CLOUDS WERE
SELECTED
3.1 Some Examples
On 27 February 2007 the director of the Informa-
tion Services Division of University College London
Are Clouds making Our Research Irrelevant and Who Is at Fault? (Position Paper)
437
(UCL) informed computer science faculty members
that: “the provost had decided (without consulting
CS), to switch to Hotmail for e-mail, but using a fake
university address. (Microsoft already had demon-
strated how to use Hotmail at some universities, e.g.,
in Australia earlier on.) The reasons that were given
for this switch was that 2 men-year would be saved
and that it was free for the first years. Whether this
really resulted in some savings was never analyzed,
so far the author knows. In fact a lot of extra peo-
ple power was needed to make the switch, making
the short term saving doubtful. One of the faculty
members pointed out that one person pushing for the
switch had a large extra income by being consultant
for Microsoft. After UCL switched, Microsoft was
able to convince other universities to stop having their
own e-mail servers. We give further details, related to
privacy, in Section 4.1.
We now give another example. To avoid the ex-
pense of landlines the University of Texas System de-
cided roughly 10 years ago to switch to Voice Over
IP (VOIP). Landline phones were removed from most
offices and replaced by ethernet connected phones
that were using TCP/IP. However, the VOIP contract
ended in 2020, and all VOIP phones were removed
from offices, leaving no phones in these whatsoever.
As solution the cloud was used. Microsoft Teams was
selected because Zoom had security problems. (For
example, Zoom originally used ECB mode, which
gives very poor encryption, in particular when en-
crypting pictures
5
.) Note that the University of Texas
System also uses Microsoft for e-mail! So, from a
privacy viewpoint, Microsoft now is in a position to
eavesdrop both “phone” conversations and e-mails!
3.2 Impact of COVID
To limit the impact of COVID lockdowns on meet-
ings, teaching, etc., the world became much more vir-
tual. CIOs decided that clouds were the fastest ap-
proach to make this virtual world a reality. Again, a
non-cloud approach was not available, so far the au-
thor knows.
3.3 Abolishment of Paper
Many organizations today have dramatically reduced
the use of paper in their administration. The solution
CIOs have used is again the use of clouds. Some star-
tups, established cloud servers, etc., have taken into
account the hierarchical structure that exists in some
organization. We describe two such examples. In a
5
https://medium.com/@TalBeerySec/zooming-on-zo
om-5-encryption-cc7e9b710b9f
typical hierarchy, annual reports are required. Role-
Based Access Control (see e.g., (Jajodia et al., 1997;
Oh and Park, 2003; Joshi et al., 2005; Sandhu et al.,
2006; Fadhel et al., 2015)) can then be used to de-
cide who should have access to these reports. An-
other example, is the chain required to receive per-
mission (e.g., to buy new equipment). Role-Based
Access Control can be used to achieve this chain of
approval.
4 CLOUD SECURITY & PRIVACY
4.1 Do CIOs Take Security and Privacy
into Account?
We start by giving two concrete examples. First we
go back to the example of the switch to the use of Mi-
crosoft’s hotmail at UCL (see Section 3.1). Privacy
concerns were almost completely ignored. The main
exception was that Google’s Gmail was not selected,
because they could not guarantee that the data would
be stored in the EU (this predates Brexit). Note how-
ever, that due to the US CLOUD Act (March 2018),
Microsoft may have to turn “overseas” data to the US
Department of Justice anyway!
Our second example is related to universities,
particularly in the US. There the Family Educa-
tional Rights and Privacy Act of 1974 (FERPA), as
amended, protects student data, such as the name
of the student. Tools used in many US universi-
ties today, such as Microsoft Teams do not allow to
hide the participants at the meeting (Microsoft Teams,
2022). Similar regulations may exist in other coun-
tries/regions and might be violated systematically to-
day.
4.2 The Impact of Closed Source
Since the software on most cloud servers is closed
source, there are a lot of things we can not say. For ex-
ample in Section 3.3 we spoke about the potential use
of Role-Based Access Control. However, since the
source and the design of the software used is closed,
we can not be certain that Role-Based Access Control
was actually used, or whether a reinvention of some
variant was used.
Since the source is closed, we can not evaluate the
software on vulnerabilities, security and privacy prop-
erties we would expect. A concrete example is given
in Section 6.
SECRYPT 2022 - 19th International Conference on Security and Cryptography
438
4.3 The Impact of Laws
In many countries attempting to hack a computer,
such as a server is illegal. This restricts researchers
to study which cloud servers are vulnerable to Denial
of Service. Moreover there is a lack of regulations for
these who design clouds. So, we end up in a world
for which we no longer can estimate how vulnerable
it is to cyber attacks. It is well known today that most
superpowers have military hacking teams, but smaller
countries such as Iran and Israel are also assumed to
have these capabilities. Not knowing how vulnerable
our society is today, is far from ideal.
4.4 Ignoring Privacy
When using clouds instead of distributed systems, pri-
vacy has already been undermined. However, the sit-
uation is worse than it seems. In many applications,
participants should not know the names of other par-
ticipants. One example is in the military context,
when people are working on a classified project. We
gave another example in Section 4.1.
Moreover, to cope with the COVID pandemic,
Chief Information Officer were sometimes forced to
choose among servers who each blatantly ignore pri-
vacy rules.
Browser security is very problematic (see
e.g. (Louw et al., 2008)) and many cloud providers
use these as user interface.
5 OUR POSITIONS
These are:
Clouds have lowered our information security.
CIOs usually do not compare competing clouds
on their security properties and do not ask inde-
pendent experts for their advice.
Self-acclaimed experts often lack basic knowl-
edge.
Research on information security is becoming ir-
relevant.
We now justify some of these positions in the next
section.
6 A CASE STUDY
The section is based on the author’s 2022
poster (Desmedt, 2022). This poster compared
the state-of-the-art on secure e-voting with a cloud
based approach. We will first briefly survey the
state-of-the-art in research on secure e-voting. Then
we discuss a cloud based approach. We then explain
how the case study inspired our positions.
6.1 State-of-the-Art Research
The research on e-voting started with Chaum propos-
ing MIX servers (Chaum, 1981). For a while the
topic was hot and many papers were published, i.e.,
too many to do a proper survey. So, therefore we
only mention what security properties can be achieved
(some requiring unproven assumptions).
Theoretical research proposed voting systems that
guarantee:
security against double voting. This property is
achieved by having the voter digitally sign the bal-
lot. This signature is then checked and removed
from the ballot.
anonymity/privacy
6
of the vote. This can be
achieved:
conditionally: MIX servers are used that per-
mute and re-encrypt all ballots (Park et al.,
1994). The ballots are encrypted by the voter.
At the end threshold decryption (Desmedt,
1994) is used to decrypt the mixed ballot (Abe,
1998).
unconditionally: in Code Voting (Chaum,
2001) to each (voter, candidate) corresponds a
unique random number. These are mechani-
cally permuted and then securely send (e.g., via
postal mail) to the voter. An anonymous chan-
nel is used by the voter to send in the vote.
correctness: zero-knowledge interactive proofs
(Goldwasser et al., 1989) are usually used to prove
that the MIX server (see e.g., (Sako and Kilian,
1994)) did not introduce errors. In homomorphic
voting, voters will use such proofs to demonstrate
correctness of the range of their vote.
no need for a trusted security agent. This prop-
erty is achieved by assuming the number of un-
trusted parties is bounded. (Some initial proto-
cols were described without taking into account
that some of the aforementioned parties may con-
spire.)
being hacking-free. Chaum’s code voting and
variants allow to vote on a computer which has
been hacked. Correctness and privacy are unaf-
fected.
6
The crypto community uses anonymity, while the pop-
ular press talks about privacy.
Are Clouds making Our Research Irrelevant and Who Is at Fault? (Position Paper)
439
universal verifiability: it allows anyone in the
world to check the correctness. This can be
achieved when replacing interactive protocols by
non-interactive ones, e.g., using the Fiat-Shamir
trick (Fiat and Shamir, 1987). The security is only
conditional.
Note that many of the aforementioned properties can
be obtained together. Usually, conditional and uncon-
ditional security are mutual exclusive, but there are
exceptions to that rule of thumb.
Note that practical problems remain, e.g.:
Some practical systems use the WWW, but the
WWW and browsers are insecure. For an ex-
ample of a voting scheme that was hacked using
a browser rootkit, see (Estehghari and Desmedt,
2010).
Some systems are not user-friendly. Chaum’s
code voting has often been mentioned in this con-
text, but variants were proposed that allow more
classical voting (see e.g., (Desmedt and Erotokri-
tou, 2015)).
6.2 A Cloud based Approach
We give an example of a particular cloud implemen-
tation of e-voting and we focus in particular on a real
life scenario.
On April 2, 2021, at 6:32pm CDT, the Acting
Head of Computer Science at University of Texas at
Dallas sent an e-mail that: 48 votes were received
from 40 eligible voters! So, what went wrong?
Earlier, at 2:34 pm CDT the same day, the author
informed the Acting Head that “if I go to a Hotspot
which changes my IP address (and I also change my
MAC address), I can vote a 2nd time!!” (Later some-
one else observed that you can just vote twice from
the same IP address!)
The first solution that was proposed is to have
someone collect votes in the clear (i.e., violate pri-
vacy). A deeper analysis, using a Google search,
found that at Tufts University they explain the differ-
ence between:
a “Survey Link” (which does not prevent repeat
voting), and
“The Qualtrics Mailer,requiring an individual e-
mail is sent to each authorized voter.
The last approach was eventually used to vote for the
candidate to hire.
6.3 A Comparison
First of all it is clear that Qualtrics’ voting system has
a major usability problem! Indeed, anyone using it
should immediately know that when using the Survey
Link, that double voting is not prevented.
We now compare the Qualtrics’ voting system
with the state-of-the-art in e-voting, which we briefly
surveyed in Section 6.1. We organize this secu-
rity comparison into two categories, being these for
which:
we can not state anything. The black box and
closed source approach that seem to have been
used implies that:
we do not know whether these who imple-
mented this voting system took the desired pri-
vacy and security properties of e-voting into ac-
count.
we also can not answer whether the designers
are aware of the state-of-the-art in the area of
voting.
This implies in particular that after consulting
Qualtrics’ documentation
7
one can still not an-
swer the following questions:
What privacy (anonymity) guarantees does
Qualtrics offer?
Will Qualtrics keep the votes for eternity?
What about guaranteeing correctness?
How secure is Qualtrics server against hacking
(e.g., how easy is it to double vote)?
What mechanisms do they use to guarantee the
aforementioned?
Has Qualtrics been certified?
What NIST standards does Qualtrics follow?
can compare with the state of the art. In this
case, such a comparison is not possible and we are
even unable to state which privacy/security prop-
erties The Qualtrics Mailer lacks.
6.4 How Was This Cloud Server
Chosen?
Seeing the many unanswered security questions, one
can wonder how the University of Texas at Dallas
(and other organizations) decided to use Qualtrics
Mailer. It poses serious questions how clouds are cho-
sen. Indeed:
As the author already stated during his NIST talk
on June 7, 2011:
One can wonder whether CEO’s have their
head in the cloud instead of both feet on the
ground, when rushing to adopt cloud tech-
nology?
7
https://www.qualtrics.com/marketplace/vote-rank-su
rvey/
SECRYPT 2022 - 19th International Conference on Security and Cryptography
440
Is the Chief Information Officer the sole person to
decide what clouds to use, or is there an advisory
board?
Are clouds evaluated by technical experts?
6.5 How e-Voting Inspired our Positions
The huge difference with what the research on e-
voting can offer, which seem not to be a part of
Qualtrics Mailer, is likely just one example in the con-
text of what the cloud offers for privacy/security com-
pared to what it could offer. Due to the fact that many
cloud servers use a closed source and a closed de-
sign approach, many of the comparison in Section 6.3
seem to extend in a context different from e-voting.
The case study, together with the two examples
in Section 3.1, made us reflect on how we came to
the situation we are in now. These together with Sec-
tions 3 and Sections 4 were instrumental in formulat-
ing our positions, as stated in Section 5.
7 WHO IS AT FAULT?
7.1 Research Failures
Today, research on information security is often not
driven by a particular need, but by what is in fash-
ion. We now give one concrete example. Nigel Smart,
at Eurocrypt 2017, during his invited talk, pointed
out that Secure Multiparty Computation (SMC) (see
e.g., (Yao, 1986; Goldreich et al., 1991)) has no key
application.
Today most reviewers will reject papers in which
old attacks are used against new software. This might
be good from a theoretical research viewpoint, but it
clearly undermines the impact of research on the real
world! In such a research environment, it is not attrac-
tive for researchers to analyze the security of clouds!
Maybe we should be talking to engineers who have
a different viewpoint on the importance of practical
oriented research.
We now look at the role of funding agencies. First,
in-fashion topics will receive a lot of funding, and this
in different regions of the world, regardless whether
there is a need for that research. Second, one needs a
major initiative for a cloud-free distributed computer
world based on the state-of-the-art in information se-
curity. Examples of such initiatives but in other con-
texts, were NASA and the TGV. In our context, large
cloud servers would probably lobby against such an
initiative.
7.2 Researchers Setting Bad Examples
Many researchers work in a small subarea of the huge
field of information security. So, often they are not
aware of the best practices. We just give two exam-
ples, without mentioning any names.
During a 2003 Summer School, one of the orga-
nizers, who teaches Computer Security, logged in to
the university’s computer displaying the login screen
to all participants. It showed ******, implying that
the password is only 6 characters! When this was
pointed out, the person replied: “My password is hard
to guess.
Today many people working in information secu-
rity when having to copy a file from one laptop to
another, just upload it to the cloud and then download
it to the other computer, clearly showing they do not
understand the privacy violations of their actions.
7.3 Fake Experts
A question we asked is whether CIOs are using ex-
perts before making their decision. However, some
expertise is fake. As an example, some invited “ex-
pert” speakers confuse bitcoin with blockchain. Ob-
viously, self-certified experts aggravate the problem.
7.4 Lack of Regulations
Many of today’s clouds are large corporations who
have been very successful in lobbying against any
type of regulation. They will probably continue this
practice.
7.5 Customer Reviews
Some countries have magazines such as Consumer
Reports
8
, which for example evaluates cars, home ap-
pliances, etc. So far we know there is no such maga-
zine that evaluates clouds on information security as-
pects.
8 CONCLUSIONS
It seems that cloud severs were designed in a black
box way and that they used a closed source approach.
This implies that when looking at the state-of-the-art
security properties, we may be unable to state whether
a cloud server achieves these or not.
8
See: https://www.consumerreports.org/
Are Clouds making Our Research Irrelevant and Who Is at Fault? (Position Paper)
441
CIOs decided which cloud severs and services are
being used today inside an organization. In that con-
text it seems that for cloud servers having a good sales
person is more important than having experts know-
ing the state-of-the-art in privacy/security. In this con-
text, one can only conclude that:
clouds are making our research irrelevant, and that
technology transfer from research to practice has
been a failure.
ACKNOWLEDGEMENTS
Travel made possible by the Jonsson Endowment.
REFERENCES
Abe, M. (1998). Universally verifiable Mix-net with ver-
ification work independent of the number of Mix-
centers. In Nyberg, K., editor, Advances in Cryptology
Eurocrypt ’98, Proceedings (Lecture Notes in Com-
puter Science 1403), pages 437–447. Springer-Verlag.
Espoo, Finland, May 31–June 4.
Chaum, D. (1981). Untraceable electronic mail, return ad-
dresses, and digital pseudonyms. Commun. ACM,
24(2):84–88.
Chaum, D. (2001). Surevote: Technical overview. In Pro-
ceedings of the Workshop on Trustworthy Elections
(WOTE ’01), August 2001.
Chaum, D., Rivest, R. L., and Sherman, A. T., ed-
itors (1983). Advances in Cryptology. Proc. of
CRYPTO’82. Plenum Press N.Y.
Desmedt, Y. (2022). Can we trust cloud voting? Poster
at Finanicial Cryptography 2022 (not in the proceed-
ings).
Desmedt, Y. and Erotokritou, S. (2015). Making code vot-
ing secure against insider threats using uncondition-
ally secure MIX schemes and human PSMT protocols.
In Vote ID, volume 9269 of Lecture Notes in Computer
Science. Springer.
Desmedt, Y. G. (1994). Threshold cryptography. European
Trans. on Telecommunications, 5(4):449–457. (In-
vited paper).
Estehghari, S. and Desmedt, Y. (2010). Exploiting the client
vulnerabilities in internet e-voting systems: Hacking
Helios 2.0 as an example. In 2010 Electronic Voting
Technology Workshop/Workshop on Trustworthy Elec-
tions (EVT/WOTE ’10), August 9–10, 2010.
Fadhel, A. B., Bianculli, D., and Briand, L. (2015). A com-
prehensive modeling framework for role-based access
control policies. Journal of Systems and Software.
Fiat, A. and Shamir, A. (1987). How to prove your-
self: Practical solutions to identification and signature
problems. In Odlyzko, A., editor, Advances in Cryp-
tology, Proc. of Crypto ’86 (Lecture Notes in Com-
puter Science 263), pages 186–194. Springer-Verlag.
Santa Barbara, California, U. S. A., August 11–15.
Goldreich, O., Micali, S., and Wigderson, A. (1991). Proofs
that yield nothing but their validity or all languages in
NP have zero-knowledge proof systems. Journal of
the ACM, 38(1):691–729.
Goldwasser, S., Micali, S., and Rackoff, C. (1989). The
knowledge complexity of interactive proof systems.
SIAM J. Comput., 18(1):186–208.
Jajodia, S., Samarati, P., Subrahmanian, V., and Bertino, E.
(1997). A unified framework for enforcing multiple
access control policies. In SIGMOD’97, pages 474–
485. Tucson, AZ.
Joshi, J. B., Bertino, E., Latif, U., and Ghafoor, A. (2005). A
generalized temporal role-based access control model.
Knowledge and Data Engineering, IEEE Transactions
on, 17(1):4–23.
Louw, M. T., Lim, J. S., and Venkatakrishnan, V. N. (2008).
Enhancing web browser security against malware ex-
tensions. J. Comput. Virol., 4(3):179–195.
Microsoft Teams (2022). Teams - live event / webinar -
hide participants. https://answers.microsoft.com/en-
us/msteams/forum/all/teams-live-event-webinar-
hide-participants/27a824d3-4488-47ac-8d33-
eb5c6bae02ad.
Oh, S. and Park, S. (2003). Task–role-based access control
model. Information systems, 28(6):533–562.
Park, C., Itoh, K., and Kurosawa, K. (1994). All/nothing
election scheme and anonymous channel. In Helle-
seth, T., editor, Advances in Cryptology Eu-
rocrypt ’93, Proceedings (Lecture Notes in Com-
puter Science 765), pages 248–259. Springer-Verlag.
Lofthus, Norway, May, 1993.
Sako, K. and Kilian, J. (1994). Secure voting using partially
compatible homomorphisms. In Desmedt, Y. G., ed-
itor, Advances in Cryptology — Crypto ’94, Proceed-
ings (Lecture Notes in Computer Science 839), pages
411–424. Springer-Verlag. Santa Barbara, California,
U.S.A., August 22–25.
Sandhu, R., Ranganathan, K., and Zhang, X. (2006). Secure
information sharing enabled by trusted computing and
pei models. In Proceedings of the 2006 ACM Sympo-
sium on Information, computer and communications
security, pages 2–12. ACM.
Yao, A. C. (1986). How to generate and exchange secrets. In
27th Annual Symp. on Foundations of Computer Sci-
ence (FOCS), pages 162–167. IEEE Computer Soci-
ety Press. Toronto, Ontario, Canada, October 27–29,
1986.
SECRYPT 2022 - 19th International Conference on Security and Cryptography
442