Secure Keying Scheme for Network Slicing in 5G Architecture

Abstract

Network slicing is one of the key enabling technologies of evolving fifth generation (5G) mobile communication that fulfills multitudes of service demands of 5G networks. Although the concept of network slicing, its deployment scenarios and some security aspects like slice isolation are discussed in detail, key management for network slicing based applications is still not a well-investigated research area. In this paper, we propose a secure keying scheme that is suitable for network slicing architecture when the slices are accessed by the third party applications. Since the secure keying scheme is designed using a multi-party computation mechanism, it ensures the consent of monitored use cases or devices which the data is acquired. We discuss the performance, scalability and security properties of the keying scheme to demonstrate its appropriateness under evolving 5G paradigm.

Full text

Secure Keying Scheme for Network Slicing in 5G
Architecture
Pawani Porambage∗, Yoan Miche†, Aapo Kalliola†, Madhusanka Liyanage∗‡, Mika Ylianttila∗
∗Centre for Wireless Communications, University of Oulu, Finland
†Nokia Bell Labs, Espoo, Finland
‡School of Computer Science, University College Dublin, Ireland
Email: ∗[firstname.lastname]@oulu.fi, †[firstname.lastname]@nokia-bell-labs.com, ‡madhusanka@ucd.ie
Abstract—Network slicing is one of the key enabling technolo-
gies of evolving fifth generation (5G) mobile communication that
fulfills multitudes of service demands of 5G networks. Although
the concept of network slicing, its deployment scenarios and some
security aspects like slice isolation are discussed in detail, key
management for network slicing based applications is still not a
well-investigated research area. In this paper, we propose a secure
keying scheme that is suitable for network slicing architecture
when the slices are accessed by the third party applications.
Since the secure keying scheme is designed using a multi-party
computation mechanism, it ensures the consent of monitored use
cases or devices which the data is acquired. We discuss the
performance, scalability and security properties of the keying
scheme to demonstrate its appropriateness under evolving 5G
paradigm.
Index Terms—5G, Network Slicing, Security, Key Manage-
ment, Multi-Party Computation, Scalability
I. INTRODUCTION
Fifth generation (5G) of mobile communication are ex-
pected to meet the evolution in terms of capacity, performance
and spectrum access to radio-network segments [1]. More
importantly, 5G will be an evolution of extreme flexibility
and high programmability conversion in all non-radio network
infrastructure. 5G networks will serve a multitude of use cases
that have very diverse characteristics and requirements. The
most mentioned 5G use cases are categorized as enhanced
Mobile Broadband (eMBB), Ultra Reliable Low Latency Com-
munication (URLLC), and massive Machine Type Communi-
cation (mMTC) [2]. To satisfy different needs of these use
cases, network slicing is introduced in 5G that creates multiple
logical networks over a single physical network [3].
By definition, one network slice can accommodate a sep-
arate set of network functions without interfering to support
given application services [4]. In one vertical industry there
can be many standalone or combined use cases running simul-
taneously. Each of these horizontal use cases can be served by
one or multiple network slices. For instance, a smart factory
environment may accommodate all types of traffic classes
including eMBB (human users, Augmented Reality (AR) and
Virtual Reality (VR) applications), URLLC (cobots, automated
machineries), mMTC (sensor networks, IoTs) [5]. Different
traffic classes can be served by one or many network slices. In
these massive industrial verticals, when it is required to detect
an abnormality from a given data set that has a global impact
on the entire industry (i.e., smart factory), the process can be
outsourced to a third party monitoring application. Under such
circumstances, the monitoring application may have to access
one or many network slices (i.e., the corresponding network
storage resources) with the consent of the targeted use cases
or the devices.
In this paper, we propose a keying scheme to securely access
data from the network slices by the third party monitoring
applications with the consent of the monitored use cases or
devices. We discuss the retrieval of data and key management
with the help of multi-party computation mechanisms. The
performance of the proposed scheme is analyzed based on
the involvement of the network devices as an cost analysis.
Moreover, we mention the possibilities of scaling up the
keying scheme for large network and discuss the security
properties of the keying scheme.
The remainder of this paper is organized as follows: Sec-
tion II provides the background and related work on network
slicing and 5G service based architecture. Section III describes
the assumptions, the threat model and the use case. Section IV
presents the proposed key management scheme and its appli-
cability in the 5G architecture. Section V and VI respectively
provide performance and security analysis of the proposed
solution. Finally, section VII summarizes the work and draws
the conclusions.
II. BACKGROU ND A ND RE LATE D WOR K
5G core network is proposed to deploy in two phases [6]:
well-known point-to-point connections which is similar to the
current 4G-LTE (Long Term Evolution) architecture; service
based architecture (SBA). According to the definition released
by 3rd Generation Partnership Project (3GPP), the 5G SBA de-
livers the control plane functionality and common data repos-
itories by a set of interconnected Network Functions (NFs).
Therefore, SBA is more appropriate for the new cloud-native
networking models and has higher flexibility for an iterative
development process. 5G SBA is defined with major functional
elements (or network functions) and their connecting high-
level interfaces [1].
Due to its ability to support multiple service requirements
over a common network, network slicing is considered as a

key commercial driver for 5G [3]. By definition, a network
slice controls its own packet forwarding from the user end
to the cloud servers in the core network [4]. The end-to-
end slicing architecture has three segments including access
slices, core network slices, and pairing functions that connect
former two. CN (Core Network) slices are designed with the
logical separation between control and user plane functions
and the corresponding NFs such as Access and Mobility
Management function (AMF), Session Management function
(SMF), User plane function (UPF), Policy Control Function
(PCF), Authentication Server Function (AUSF), Unified Data
Management (UDM) and Network Slice Selection Function
(NSSF).
Security considerations in network slicing are associated
with many networking and communication technologies and
addressed with different perspectives [7], [8]. Strong slice
isolation is a main requirement to mitigate the spreading
of security threats among multiple network slices. When
two slices are sharing common resources, an attacker who
reveals the cryptographic materials on one slice can exploit to
affect the security functions running on another slice. Other
security considerations are addressed on inter-networking slice
communication, operations of network slice manager, slice
heterogeneity, authentication of network slice instances, and
key management.
Although slicing security has gained a high attention, not
many works are published during the recent past. The majority
of the work consider authentication and key management pro-
tocols from the user end or for the inter-slice communication.
The paper [9] proposes a cross-authentication scheme for 5G
heterogeneous networks by combining non-cryptographic and
cryptographic algorithms. In [10], a secure service oriented
authentication framework is presented to support slicing and
fog computing for 5G-enabled IoT services. It guarantees the
secure access of IoT services and privacy-preserving slice
selection. The solution includes a service-oriented three-party
key agreement to negotiate keys among IoT servers, local
fog servers, and users, based on Diffie-Hellman key agree-
ment. In [11], two heterogeneous signcryption schemes are
proposed to achieve mutual communications among network
slices deployed in different public key cryptosystems (i.e., e
public key infrastructure and certificateless public key cryp-
tography environment). Slice isolation, privacy, and managing
trust among different stakeholder and slices also matter the
most. Customizing slicing security by Software Defined Net-
working (SDN) using micro-segmentation is another approach
to isolate traffic flows related to different applications or
users [12].
However, none of those published articles address the secu-
rity considerations, possible attacks and mitigation techniques
related to networking slicing, when the slices are accessed by
external parties. Throughout this work, we describe how multi-
party computational algorithms can be used to implement
secure keying scheme among the network slice resources,
the served use cases (i.e., devices or users) and the external
entities. Our work is inspired by the scheme in [13], a
distributed approach with a hybrid cryptosystem that ensures
the confidentiality in a video surveillance system. According
to their setup, the recorded video-material is only available
to a subset of authorized users who will finally be able to
decrypt the videos. Similar to the scheme in [13], we also take
the advantage of using a multi-party computation algorithm
(i.e., Shamir’s secret sharing) to compute keys in our security
scheme.
III. USE C AS E AN D THR EAT MO DE L
A. Use case
To elaborate the key management protocol and its necessity,
we consider one particular vertical industry which is smart fac-
tory or Industry 4.0. Under the given industry vertical there are
multiple horizontal use cases or services running with different
requirements and specifications. For instance, a smart factory
environment includes numerous cyber-physical systems such
as cobots (collaborative robots), augmented reality (AR) ap-
plications and sensor networks. As illustrated in Figure 1, one
factory premises may accommodate different such use cases
and each can be served by single or multiple network slices.
When the factories are geographically dispersed, the horizontal
use cases can be also served by dedicated or shared network
slices. Each network slice owns logically isolated computation
and storage resources to perform data processing and storing
tasks to all the use cases that receive their services.
Fig. 1: Provision services for horizontal use cases in factories
using dedicated and shared network slices.
In addition to the aforementioned dedicated services re-
quired by different use cases, there are certain occasions
that need them to be monitored in common by Third Party
Monitoring Applications (TPMAs). For example, while de-
tecting anomalies in a particular process in the factory, the
third party application has to acquire data from multiple use
cases and keep the proper co-ordination among them and the
respective network slices. Under such a scenario, the third
party application has to access the data from the respective
network slices, however with the consent of the particular use
cases (Figure 2) or their individual components (Figure 3).
To cater such secure communication links among the third
party applications, the network slice resources, and the use

Fig. 2: Third party application access data with the consent of
different use cases.
Fig. 3: Third party application access data with the consent of
individual components in a given use case.
cases, we propose a secure key management scheme based on
Shamir’s secret sharing technique.
B. Threat model
1) Adversaries: We consider three possible adversary types.
Internal adversaries with administrative control: An inter-
nal adversary can be located in the factory premises and take
the administrative control of the devices in the factory. This
can be a tenant inside the factory or an external intruder who
has access to a compromised device.
External adversaries with access to data transportation:
An external adversary that could attack the data transportation
between TPMA, factory devices and network slice.
External adversaries with access to TPMA: An external
adversary that could attack a TPMA by accessing its security
credentials or creating Distribute Denial of Service (DDoS)
attacks.
2) Attacks on a slice while accessing by a third party:
DDoS flooding attacks: These can be launched by external
adversaries on network slices (i.e., particularly on AMF) while
TPMA communications in all cases are occurring across an
untrusted network.
Data tampering attacks: These attacks can be launched by
external adversaries when they access data obtained by TPMA.
When a TPMA is compromised the attacker can deliberately
modify, replay or inject bogus data stored. Moreover, tamper-
ing attacks can also be initiated by the external adversaries
those who can access the data transportation.
Key-compromise impersonate attacks: These attacks may
occur when an internal adversary uses a compromised device
to access its keying materials use them for future commu-
nication purposes. If the long term private key shares of the
devices are compromised, the attacker can use those key shares
to compute the security credentials requested by TPMA.
C. Assumptions
The slice operators’ network is functional as the service-
based 5G core network architecture. Within the factory
premises, the monitored devices are protected by a Trusted
Platform Module (TPM) on each device for mutual authen-
tication and integrity checks in all communications. The
communication link establishment between the TPMA and 5G
core is performed by the means of a conventional security
protocol like IPSec and TLS (Transport Layer Security).
Moreover, we consider an ideal situation where the proper
slice implementation is achieved in such a way to avoid slice
specific security vulnerabilities including slice isolation and
side channel attacks.
IV. PROPOSED KEYING SCHEME
A. Preliminaries
For all the calculations in the rest of the sections, we
consider that Gqis a group of prime order q, where the
Discrete Logarithm Problem and closely related problems are
believed to be hard and gis a generator of Gq. All the
computation in Zare also undergoing mod q, although it is not
appeared in the text. In addition to that, the proposed security
architecture exploits Shamir’s secret sharing [14] technique
to distribute and reconstruct the shares of private keys and
ElGamal cryptosystem [15] for encryption and decryption of
interval-keys. Shamir’s secret sharing technique is based on
a(n, k)threshold scheme [14], wherein ndevices process
a polynomial share and kpolynomial shares being enough
to reconstruct the DH keys through the Lagrange polynomial
interpolation. According to [14], the (n, k)threshold scheme
is selected as n= (2k−1).
B. Keying scheme
As illustrated in Figure 4, the devices {D1, D2, . . . , Dn}are
accessing one particular network slice. When a TPMA needs
to acquire data related to the given devices from the network
slice, it also has to get the consent of those devices.
As an initial phase, the network slice and the devices will
receive the corresponding cryptographic keys from a Key
Distribution Center (KDC), which is co-operating with the
Authentication Server Function (AUSF). For a given network
slice that serves ndevices (or ndistinctive use cases), KDC
generates a key-pair (d, e)for the ELGamal cryptosystem: d
and eare private and public keys. The shares of the private
key (d1, d2, . . . , dn) and the public key e=gdare respectively
delivered to the devices and the network slice in a secure
manner. The secret key dis generated following a t-degree

Fig. 4: Graphical representation of delivering security creden-
tials in the key management scheme.
polynomial f() (i.e., d=f(0)) and dishares are computed
as follows.
di=f(i) = d+
t
X
j=1
rjij, rj∈RZ∗
q(1)
According to Shamir’s Secret Sharing technique and by La-
grange interpolation, in order to reconstruct the private key d,
it is necessary to have at least t+ 1 number of divalues.
There can be multiple security and privacy issues arising
when the third party application can keep the records of
previously accessed data from a network slice. Therefore, it is
necessary to ensure that the data released by the network slice
is transferred as encoded context in an application specific
format (e.g., use homomorphic encryption). There should be
an entity operating at the network slice, that is responsible for
encoding the data and providing it for pre-designed tasks of
TPMA.
The encoded data related to a particular incident or a
given period of time is named as Mand encrypted with an
interval-key kusing symmetric encryption ES(M, k) = c. The
network slice generates a random interval-key kas a one-
time key related to the request made by TPMA. In addition
to that, the slice encrypts kwith the public key eusing
ElGamal asymmetric encryption and computes c1and c2
values: EA(e, k)=(c1, c2).
EA(e, k) = (gα, ekα)=(c1, c2)α∈RZq(2)
When TPMA accepts the received message (i.e., c, c1, c2),
it has to follow several steps to derive kand reconstruct the
encoded data. First step is to derive the ephemeral key k, using
(c1, c2)and the consent of the devices (or the use cases) in the
factory. On behalf of the TPMA, the network slice manager
sends c1value to the devices who agree to co-operate. Then
each collaborating device performs the ElGamal decryption of
c1using its private key share di(i.e., equation 3) and sends
back c1ito TPMA through the network slice manager.
DA(c1, di) = cdi
1=c1i(3)
After collecting the sufficient number of decrypted values
(i.e., from t+ 1 cooperative devices) from ndevices, TPMA
derives the interval-key kby decrypting the values as follows:
First compute λicoefficients for each received c1iin the subset
P(i.e., size of t+ 1) of the participating devices.
λi=Y
i∈P,j6=i
−j
i−j(4)
Then kvalue is decrypted using the Lagrange formula and λi
values.
D((cλ1
11 , cλ2
12 , . . . , cλ(t+1)
1(t+1)), c2) = c2(Q
i∈P
cdi×λi
1)−1
=c2(c
P
i∈P
di×λi
1)−1
=c2(cd
1)−1
=k
(5)
Having the key k, TPMA can decrypt the encoded data by
performing DS(c, k) = M.
C. High-level description of proposed scheme
This section describes the high-level overview of the infor-
mation flow of the proposed protocol under the umbrella of
next generation core network architecture (Figure 5).
Fig. 5: Information flow of the proposed keying scheme in 5G
service based architecture.
First, TPMA sends a registration request to Access and
Mobility management Function (AMF). Then AMF selects
an Authentication Server Function (AUSF) and interacts with
User Data Management (UDM), where TPMA can perform the
registration and primary authentication specified in 3GPP to
access 5G network. If the authentication process is successful,
TPMA can proceed with sending service request and authen-
ticating the service to AUSF. Based on the service request,
AUSF generates the ephemeral key kfor the session, and
shares it with AMF. Then AMF selects a network slice instance
based on TPMA’s service request and sends session estab-
lishment request to the corresponding Session Management
Function (SMF) of the slice along with the key k. The slice
specific function, User Plane Function (UPF) manages user
plane traffic. Moreover, the network slice receives the service
area restrictions of the serving devices and the accessibility of

the stored data from UDM. Based on the accessibility policies,
the slice itself accesses the encoded data M, computes c, c1, c2
values and sends c1to the devices. The devices that are willing
to cooperate compute c1ivalues and sends back to the network
slice. Finally, the network slice will send c, c1, c2and all the
received c1ivalues. Since there is no direct communication
between the devices and TPMA, c1ivalues are delivered via
the network slice manager.
V. PERFORMANCE ANALYSIS
To discuss the performance of the proposed keying scheme,
we considered the behaviour of the devices in the factory
premises in terms of their willingness to cooperate with TPMA
to reconstruct the secrets. TPMA requests rightful access to
the data, in order to provide an additional service or to detect
an abnormality related to the devices. Therefore, if the devices
be cooperative with TPMA each device will receive a gain (G)
in return. At the same time, while contributing for the secret
reconstruction, each device has to undertake a cost (C) that
includes computation and communications costs.
For the successful key reconstruction, at least there should
be t+ 1 cooperative devices. Consequently, the total profit
which one device can gain will also depend on the likelihood
of the other devices to cooperate with key reconstruction. We
consider that each device is willing to cooperate (i.e., Di= 1)
with an equal probability of a: the probability that a device
will not cooperate is 1−a(i.e., Di= 0). Then the probability
of success or the probability that at least t+ 1 devices are
contributing among ndevices (i.e., D1, ...Di, ...Dn) will be:
P rsuccess =P r{
n
P
i=1
Di≥(t+ 1)}
=n
t+1a(t+1)(1 −a)(n−(t+1)) +· · · +n
nan
=
n−1
P
i=tn
i+1a(i+1)(1 −a)(n−(i+1))
(6)
Therefore the expected net profit (Pavg ) at a device would be
the difference between its expected net gain and the expected
cost:
Pavg =G
n−1
X
i=tn
i+ 1a(i+1)(1 −a)(n−(i+1)) −C a (7)
The graphs in Figure 6 show the behaviour of the expected
net profit (Pavg ) with the variation of the probability of device
cooperation (a). We keep a fixed gain value as G= 20
and changed the cost value C. Thereby we observe that
Pavg is increasing with the decremented C. In each case, the
maximum Pavg is observed for the probability of abetween
0.8and 0.9.
A clustering mechanism with a hierarchical key distribution
scheme can be used to tailor the proposed keying scheme in
such a way to provide higher scalability for the given use case.
As described in Section III, network slices can be assigned
under multiple scenarios (i.e., dedicated or shared slices) or
TPMA may request data from the devices served by different
slices. Moreover, when the number of devices served by one
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Device cooperating probability (a)
-10
-5
0
5
10
15
20
Expected net profit (Pavg)
C = 0.1
C = 1
C = 5
C = 10
C = 20
Fig. 6: The behaviour of the expected net profit with the
probability of device cooperation at different costs
network slice drastically increases from hundreds to thousands
(e.g., sensor networks), it will be challenging to adjust the
values of nand taccordingly. One responsible node (e.g.,
cluster head) can be assigned to a group of devices and each
of those cluster heads can provide the consent to the TPMA
on behalf of their leaf nodes (i.e., end devices).
VI. SE CU RI TY P ROP ERT IE S
Security properties of the keying scheme are described
considering the threat model mentioned in Section III. The
keying scheme takes the advantage of multitude secret shares
to reconstruct key kgenerated by AUSF. Since the scheme
takes the consent of the devices (or use cases), it can be easily
adopted for a group of users and the protocol may provide
an implicit assurance for privacy-protection of the monitored
users.
1) DDoS flooding attacks: Although the devices provide
their consents to the key reconstruction process, they do not
have direct communication with TPMA. Instead the message
is coming from the slice itself, which is assumed to deliver
securely. Therefore, on the networking devices, the protocol
will clearly prevent the DoS attacks created by the external
parties. However, since all the TPMA communications are
occurring across an untrusted network, it is necessary to
apply some conventional protection mechanisms such as IPSec
and TLS. Standard DDoS defensive mechanisms should be
implemented at the network slice manager which is the first
contact point of the TPMA [16].
2) Data tampering attacks: When there is a request coming
from a TPMA, the slice will always share the encoded data
which are associated with that stand-alone request. Since the
data is in encoded form, it will be hard to an external attacker
to retrieve any useful information out of it for future attacks.
Moreover, sending encrypted data will prevent the tampering
attacks during the data transportation.
3) Key-compromise impersonate attacks: According to the
Lagrange polynomial interpolation t+ 1 number of private
key shares out of nshares are used to reconstruct the key
k. Therefore, as long as an internal adversary is not able to
solve discrete logarithm problem and do not compromise at

least t+ 1 devices in the factory, he is not able to get any
information about the private key (d) and reconstruct the key
k. Regular updates of key shares and the polynomials will also
increase the security strength of the scheme and mitigate the
risk of key-compromise attacks.
VII. CONCLUSION
In summary, we proposed a key scheme for network slicing
based systems that provide secure accessibility for third party
monitoring applications with the consent of the networking
devices. The keying scheme is discussed considering its perfor-
mance, scalability and security properties along with its high-
level integration in 5G service based architecture. According
to the performance analysis of the keying scheme, the devices
will receive the maximum average net profit on a cooperation
probability between 0.8 to 0.9. The proposed solution has great
flexibility to tailor its behavior and characteristics based on the
required security strength and use case scenario. Our future
research is focused on implementing the proposed keying
scheme and analyzing its behaviour with respect to the delay
and scalability.
ACKNOWLEDGMENT
This work has been performed under the framework of
6Genesis Flagship (grant 318927), 5GEAR, SecureConnect
and RESPONSE 5G (Grant No: 789658) projects. This re-
search is funded by the Academy of Finland, Business Finland
and the European Union.
REFERENCES
[1] 3GPP, “System Architecture for the 5G Systems,” Technical Specifica-
tion, June 2018. [Online]. Available: https://www.etsi.org/deliver/etsi
ts/123500 123599/123501/15.02.00 60/ts 123501v150200p.pdf
[2] A. Gupta and R. K. Jha, “A survey of 5g network: Architecture and
emerging technologies,” IEEE access, vol. 3, pp. 1206–1232, 2015.
[3] 3GPP, “Study on management and orchestration of network
slicing for next generation network,” Technical Specification,
June 2018. [Online]. Available: https://portal.3gpp.org/desktopmodules/
Specifications/SpecificationDetails.aspx?specificationId=3091
[4] International Telecommunication Union, “Terms and definitions for
IMT-2020 network: ITU-T Y.3100 (09/2017),” Sep 2017.
[5] Y. Siriwardhana, P. Porambage, M. Liyanage, J. S. Walia,
M. Matinmikko-Blue, and M. Ylianttila, “Micro-Operator driven
Local 5G Network Architecture for Industrial Internet,” in IEEE
Wireless Communications and Networking Conference (WCNC).
IEEE, 2019, pp. 1–8.
[6] I. Afolabi, T. Taleb, K. Samdanis, A. Ksentini, and H. Flinck, “Network
slicing and softwarization: A survey on principles, enabling technolo-
gies, and solutions,” IEEE Communications Surveys & Tutorials, vol. 20,
no. 3, pp. 2429–2453, 2018.
[7] NGMN Alliance, “5G security recommendations package 2: Network
slicing,” Apr 2016.
[8] R. Khan, P. Kumar, D. N. K. Jayakody, and M. Liyanage, “A Survey on
Security and Privacy of 5G Technologies: Potential Solutions, Recent
Advancements and Future Directions,” IEEE Communications Surveys
& Tutorials, 2019.
[9] C. M. Moreira, G. Kaddoum, and E. Bou-Harb, “Cross-layer authenti-
cation protocol design for ultra-dense 5g hetnets,” in IEEE International
Conference on Communications (ICC), 2018, pp. 1–7.
[10] J. Ni, X. Lin, and X. S. Shen, “Efficient and secure service-oriented
authentication supporting network slicing for 5g-enabled iot,” IEEE
Journal on Selected Areas in Communications, vol. 36, no. 3, pp. 644–
657, 2018.
[11] J. Liu, L. Zhang, R. Sun, X. Du, and M. Guizani, “Mutual heterogeneous
signcryption schemes for 5g network slicings,” IEEE Access, vol. 6, pp.
7854–7863, 2018.
[12] J. Suomalainen, K. Ahola, M. Majanen, O. M¨
ammel¨
a, and P. Ruuska,
“Security Awareness in Software-Defined Multi-Domain 5G Networks,”
Future Internet, vol. 10, no. 3, p. 27, 2018.
[13] M. Schaffer and P. Schartner, “Video Surveillance: A Distributed
Approach to Protect Privacy,” in Communications and Multimedia
Security, ser. Lecture Notes in Computer Science. Springer Berlin
Heidelberg, 2005, vol. 3677, pp. 140–149. [Online]. Available:
http://dx.doi.org/10.1007/11552055 14
[14] A. Shamir, “How to Share a Secret,” Communication ACM, vol. 22,
no. 11, pp. 612–613, 1979.
[15] T. El Gamal, “A Public Key Cryptosystem and a Signature Scheme
Based on Discrete Logarithms,” in Proceedings of CRYPTO 84 on
Advances in Cryptology, 1985, pp. 10–18.
[16] J. Mirkovic and P. Reiher, “A taxonomy of ddos attack and ddos
defense mechanisms,” ACM SIGCOMM Computer Communication Re-
view, vol. 34, no. 2, pp. 39–53, 2004.