centralised e-mail service on their devices, it is pos-
sible to devise next-generation, Net-Rat-aware gate-
ways without having to rely on legacy infrastructure.
An important consideration when transforming
applications to end-user devices are security aspects.
Firstly, users are effectively burdened with the re-
sponsibility of being service operators.This raises the
second important issue: Maintaining the integrity of
every user’s devices is imperative. Therefore, ways
of detecting Byzantine faults (Lamport et al., 1982)
also need to be considered. Approaches such as Peer-
Review (Haeberlen et al., 2007) deal with this issue,
however, the problem definition resulting from the
Net Rat’s architecture is simpler: Most importantly,
the impact of a compromised Net Rat instance is com-
parable to that of a single rogue user in a traditional,
centralised setup.
The security evaluation results illustrate how the
Net Rat achieves its security properties and how these
differ based on the chosen deployment. As our system
enforces encrypted storage and never stores authenti-
cation information, even attackers with full disk ac-
cess cannot extract any information. In combination
with a robust network layer taking care of transport
security, the Net Rat indeed provides increased secu-
rity compared to traditional services even for worst-
case scenarios. Most importantly, innocent users do
not suffer collateral damage from infrastructures out-
side their control being compromised. Furthermore,
due to all data being directly distributed among de-
vices, availability can be guaranteed to some degree
even during network downtimes.
Device enrolment has not been considered. How-
ever, leaving this issue open for future work does not
impact the properties of our system, as soon as de-
vices have been connected. In essence, enrolment is a
separate issue. When examining single, multi-device-
user scenarios, the burden of performing sophisticated
enrolment procedures does not outweigh its benefits.
Another area of future work includes supporting user
authentication methods which do not rely on a shared
secret like client certificates, or the Qualified Mobile
Server Signature (Orthacker et al., 2010).
The final area of future work concerns a more sophis-
ticated distributed storage. It must be possible to cater
to memory heterogeneity within a set of connected
devices for the Net Rat to be applicable to a wide
range. However, due to the fact that events are the
smallest unit of information, this issue is definitely
within reach. After all, event replication already lies
at the core of the Net Rat’s workflow. Therefore,
memory heterogeneity can be catered to by manag-
ing events in a clever way. By tweaking the ETDC,
this issue can be solved separately from the Net Rat’s
core functionality much like device enrolment.
The results of our case study and the outcome
of the security evaluation clearly show the feasibility
of decentralising services. This highlights how exist-
ing services can be improved through decentralisation
and presents opportunities to develop novel services
based on the solid foundation provided by the Net Rat.
REFERENCES
Blum-Dumontet, E. (2017). Defeating encryption:
the battle of governments against their people.
https://www.privacyinternational.org/node/1427, Re-
trieved: 2017-04-24.
Cohen, B. (2013). The BitTorrent Protocol Specifica-
tion. http://www.bittorrent.org/beps/bep 0003.html,
Retrieved: 2017-04-24.
Dingledine, R., Mathewson, N., and Syverson, P. (2004).
Tor: The Second-Generation Onion Router. In Pro-
ceedings of the 13th USENIX Security Symposium.
Dropbox, Inc. (2016). Dropbox Privacy Policy.
https://www.dropbox.com/privacy/, Retrieved: 2016-
09-27.
Fiadino, P., Schiavone, M., and Casas, P. (2015). Vivisect-
ing WhatsApp in Cellular Networks: Servers, Flows,
and Quality of Experience. In Traffic Monitoring and
Analysis: 7th International Workshop, pages 49–63.
Springer International Publishing, Cham.
Google, Inc. (2014). Google Terms of Service.
https://www.google.com/intl/en/policies/terms/, Re-
trieved: 2017-04-24.
Haeberlen, A., Kouznetsov, P., and Druschel, P. (2007).
PeerReview: Practical Accountability for Distributed
Systems. In Proceedings of Twenty-first ACM
SIGOPS Symposium on Operating Systems Princi-
ples, SOSP ’07, pages 175–188, New York, NY, USA.
ACM.
Hautakorpi, J., Camarillo, G., and Lopez, D. (2009). Frame-
work for Decentralizing Legacy Applications. In Pro-
ceedings of the 2009 9th IEEE/ACM International
Symposium on Cluster Computing and the Grid, pages
544–549, Washington, DC, USA. IEEE Computer So-
ciety.
International Organization for Standardization (2014).
ISO/IEC 15408-1:2008 Information technology —
Security techniques — Evaluation criteria for IT se-
curity — Part 1: Introduction and general model.
Lamport, L., Shostak, R., and Pease, M. (1982). The Byzan-
tine generals problem. ACM Transactions on Pro-
gramming Languages and Systems, 4(3):382–401.
Maymounkov, P. and Mazi
`
eres, D. (2002). Kademlia:
A peer-to-peer information system based on the xor
metric. In Revised Papers from the First Interna-
tional Workshop on Peer-to-Peer Systems, pages 53–
65, London, UK. Springer.
Orthacker, C., Centner, M., and Kittl, C. (2010). Qualified
Mobile Server Signature. In Security and Privacy –
SECRYPT 2017 - 14th International Conference on Security and Cryptography
382