Advances in Internet of Things, 2012, 2, 1-7 Published Online January 2012 (
Cyber-Physical-Social Based Security Architecture for
Future Internet of Things
Huansheng Ning, Hong Liu
School of Electronic and Information Engineering, Beihang University, Beijing, China
Received December 7, 2011; revised December 28, 2011; accepted January 15, 2012
As the Internet of Things (IoT) is emerging as an attractive paradigm, a typical IoT architecture that U2IoT (Unit IoT
and Ubiquitous IoT) model has been presented for the future IoT. Based on the U2IoT model, this paper proposes a
cyber-physical-social based security architecture (IPM) to deal with Information, Physical, and Management security
perspectives, and presents how the architectural abstractions support U2IoT model. In particular, 1) an information se-
curity model is established to describe the mapping relations among U2IoT, security layer, and security requirement, in
which social layer and additional intelligence and compatibility properties are infused into IPM; 2) physical security
referring to the external context and inherent infrastructure are inspired by artificial immune algorithms; 3) recom-
mended security strategies are suggested for social management control. The proposed IPM combining the cyber world,
physical world and human social provides constructive proposal towards the future IoT security and privacy protection.
Keywords: Internet of Things; Physical; Social; Cyber, Security; Architecture
1. Introduction
The Internet of Things (IoT) becomes an attractive research
topic, in which the real entity in physical world becomes
virtual entity in cyber world, and both physical and digital
entities are enhanced with sensing, processing, and self-
adapting capabilities to perform interaction through spe-
cial addressing scheme [1]. Along with the combination
of Internet and modern sensor technologies such as Radio
Frequency Identification (RFID), Near Field Communi-
ca tion (NFC), and Wireless Sensor and Actuator Networks
(WSAN), IoT itself is suffering from more rigorous se-
curity challenges. Several issues in terms of system archi-
tecture, standard, and human involvement are subsequent-
ly raised. The following security p roblems seem to be in-
tense speculations, such as how to design appropriate se-
cu rity framework for things’ intelligent applications? What
is advanced security technology applied into mass data
processing? How to maintain a balance between things’
high security requirements and supporting infrastructures’
hardware limitation? And how human society securely
participates in both cyber and physical worlds with inter-
Such significant obstacles influence the development
of the future IoT, along with the exposure of mass data
which causes various potential vulnerabilities from robust
adversaries. Besides, resource restrictions including het-
erogeneous networks and sensor nodes, communication
channels/interfaces, bandwidth, storage, and energy, may
al so induce unique model design. Towards the general IoT,
studies on its architecture model, standard, communica-
tion protocol, and network management have been re-
searched [2-5]. Towards the particular IoT security, there
are several open issues such as cryptographic algorithms,
authentication protocols, access control, trust/privacy, and
governance frameworks [6]. Several researches mainly fo-
cus on specific communication techniques (e.g., WLAN,
RFID) [7,8], detailed cryptographic mechanisms (e.g., key
management) [9], and practical applications (e.g., supply
chain management, multimedia traffic) [10,11]. Meanwhile,
the security frameworks in traditional networks can also
provide merits for IoT security protection. However, se-
curity issue towards the fu ture IoT is no t a simple techni-
cally tough problem, but a multidimensional topic which
co mbines the information security, network security, infras-
tructure security, and management security. Most existent
schemes provide solutions for special communication te-
chniques or applicatio ns, which may lack universality for
the complicated system. Thus, we will establish an inte-
grated security architecture to promote universal security
consideration for the future IoT.
In the paper, we focus on a typical future IoT archi-
tecture (short for U2IoT) which comprises two subsys-
tems that Unit IoT and Ubiquitous IoT [12]. In the U2IoT
model, conceptions of mankind neural system and social
organization framework are introduced for the future IoT.
Thereafter, we propose a systematic security architecture
opyright © 2012 SciRes. AIT
(named IPM) by integrating the awareness and interact-
tivity of cyber world, physical world, and human social
into the U2IoT model. Meanwhile, the proposed IPM is
presented with embedded interaction s among information,
physical, and management. Specifically, 1) information se-
curity model with the considerations for basic and advanced
security requirements that are mapped into the security
layer to deal with sensing, networking, appli cati on, and so-
ci al attribution; 2) physical security including external con-
text and inherent infrastructure are inspired by artificial
immune, and it ensures that the things should be adapt-
able to dynamic semantic contexts with innate and adap-
tive immunities against malicious attacks; 3) management
security provides recommended strategies for hierarchical
classified scenes with rationality and compatibility. IPM
rea lizes the unison of cyber world, physical world and hu-
man social to guarantee securi ty and privacy for U2IoT.
The remainder of the paper is organized as follows. In
Section 2, we illu stra te th e ex isten t U2Io T model, an d p ro -
pose the security architecture (IPM). The main features
of IPM referring to information security, physical security,
and management security are given in Section 3. Finally,
Section 4 draws a con clusion.
2. Proposed Security Architecture for U2IoT
The U2IoT model is shown in Figure 1(a), which is es-
sentially a heterogeneous system including Unit IoT and
Ubiquitous IoT. Thereinto, Unit IoT resembling human
neural network, refers to the basic cell providing solutions
for special applications. Ubiquitous IoT includes the in-
dustrial IoT, local IoT, national IoT, and global IoT which
is integration of multiple Unit IoTs with ubiquitous fea-
tures, and it is similar to the social organization framework.
Concretely, Unit IoT comprises IoT networks and sensors,
distributed control nodes, and management and centralized
data center (M&DC), and Ubiquitous IoT respectively in-
cludes iM&DC, lM&DC, and nM&DC, for the industry,
local, and national IoTs. Figure 1(b) illustrates the pro-
posed security architecture (IPM) that addresses U2IoT
security in three perspectives.
Figure 1. U2IoT model and its security architecture (IPM). (a) The U2IoT (Unit IoT and Ubiquitous IoT) model; (b) The
proposed security architecture (IPM) based on U2IoT.
Copyright © 2012 SciRes. AIT
I: Information security includes two p erspectives (i.e.,
security layer and security requirement). Awareness
of information data is captured, interpreted, and rep-
resented by things’ capability, alon g with aggregation
algorithms, protocols, and functions are included for
intelligent information interactions.
P: Physical security relates to environmental monito-
ring, motion detection, localization, tracking, perimeter
control, and consumption supervision. The concept of
artificial immunity is applied to detect passive and ac-
tive defenses for maintaining homeostasis.
M: Management security provides the recommended
application requirement, industry/local/national regu-
lation, and international policy and standard to guide
activities and events in the human so cial.
In IPM, human social activities occurring in physical
world are identified and mapped into the unique cyber
world, which realizes harmonious unification of human,
network and things. Such triple relationships of entity in
U2IoT, makes entity identification and service discovery
are effectively performed in current cyber-physical world,
and are easily to extend to human social and its social
networks. The security aspects underline main character-
istics of U2IoT entities, as shown in Table 1.
3. Proposed Security Architecture: IPM
3.1. Information Security
Information security protects both raw data and contex-
tualized information, and an information security model
that co mprises U2IoT, security layer, and security require-
ment is established in Figure 2.
1) Co nsiderin g S ocial Factor for Secu rit y Layer: U2IoT
is generally divided into four layers as follows.
Sensor layer: it comprises generalized sensors and
gateways to perform entity identification and service
discovery. The function of sensor layer is to perceive
the entities, to extract information, and to realize se-
mantic resource discovery. The sensor techniques are
applied to realize effective integration and interaction
adaptation of the collected uncertain information.
Network layer: it includes network interfaces, com-
munication channels, network management, informa-
tion maintenance, and intelligen t processing. The cen-
tralized, distributed, and hybrid network topologies
are involved to assist monitoring and maintaining the
real-time network configuration. The network layer en-
sures reliable information transmission by adopting
data coding, extraction, fusion, restructuring, mining,
and aggregation algorithms. The main function is to
transfer and process the information obtained by sen-
sor layer, and to realize data exchange among large-
scale heterogeneous networks.
Application layer: it exports functionalities for spe-
cific applications, and provides embedded interfaces for
Table 1. The main characteristics of U2IoT en tities .
Characteristics Descriptions
Cyber, Physical, Social
Any entity exists in the physical world, along
with its existence in the cyber world in the
virtual form via specific communication an
network technologies; and the entity also has
its social identity and attributionwhich no
isolated from cyber and physical attribut io n .
Connectivity, and
Any entity can interoperate and collaborate
with other heterogeneous entities within its
ac cess domain, and the entities are interrelate
and interact on each other.
Space-time Consistency
Any entity can dynamically interact with othe
entities at any time, any place, and in any mode;
the entity can freely enter/leave the networks
without influencing the ongoing
communications; and synchronization is needed
during heterogeneous network accessing.
Multi-identity Status
Any entity has multi-identity statuses that
include a unique core identity, and ot he r
temporary identities according to its
underlying applications.
Figure 2. Security layer and security requirement.
infrastructures to perform testing, monitoring, or auditin g
applications. The standard protocols and service composi-
tion technologies are applied to realize the integration
between heterogeneous distributed networks an d it s appli-
cations, such as logistics monitoring, smart grid s ched ulin g ,
intelligent search, and cloud computing . S uc h ap p li ca t io ns
should adapt in dynamic environments.
Particularly, an additional social layer on the top of the
architecture considers the social attribution in U2IoT.
The social layer is mainly d evoted to co mmunica te amon g
objects and other supporting networks to perform corre-
lation between the cyber individual and the correspond-
ing profile in social networks. Correlative social attribu-
tions are granted to each entity, and hierarchical manage-
ment and data centers operate overall security considera-
tions. In social layer, diverse interfaces are accessed by a
real entity which acts on its correspond ing cyb er entity to
control its behaviors. Meanwhile, other social composi-
Copyright © 2012 SciRes. AIT
tions are also considered, such as ownership control mana-
gement, social relationship mod eling, and entity b ehavior
In the perspective of information security, the sensor
and network layers specify mu ltiple networks and sensor
nodes, which are used to capture data streams, to detect
activities and events with identification algorithms, and to
realize specific application functions. Core of data acqui-
ring is sensor technology (e.g., RFID, WSN, femtocell)
and Global Sensor Network (GSN) middleware, whose se-
curity is challenged by constrained resources. Note that dis-
tributed control nodes provide the capabilities to survive
under formidable conditions, and by information security
controls such as error detection and correction, random
access control, and fault tolerance are recommended.
2) Adding Intelligence and Compatibility for Security
Requirement: Elements of security requirements include
CIA Triad, authority, non-repudiation, and privacy. Ad-
ditional requirements that intelligence and compatibility
are added into advanced security considerations, which
provide reliable security and privacy protection. Table 2
presents the comparison of security requirements among
the traditional network, general IoT, and U2IoT.
Intelligence represents that an entity should own ab-
stract capabilities including self-learning, self-adapting, and
self-reasoning to adapt itself to dynamic semantic envi-
ronments. In non-deterministic channels and open inter-
faces, virtual intelligent entities should be autonomously
interconnected in U2IoT. The embedded intelligence makes
the entity have strong efficacies to adapt dynamic infor-
mation that is the mapping of environment interactions,
social connections, and human behaviors.
Table 2. Comparision of security require me nts.
Requirement Traditional network [13] General IoT [14,15] Additional Requirements in U2IoT
CIA Triad
a) Data Confidentiality:
Protect data from unauthorized disclosure;
b) Data Integrity:
Ensure correctness or accuracy of data;
c) Data Availability:
Ensure that there is no denial of authorized
access to network elements, information flows,
services and applications.
Bottom-up cryptographic algorithms
should be recommended fo r
heteroge neous network i nfr astr uctur e:
- Key management infrastructures;
- Block/Stream ciphers algorithms;
- Digital signatures;
- Hash function;
- Pseudo-random number generators.
a) Ubiquitous forward, backward, and
ongoing security:
Ensure ubiquitous unlinkability among
the prior, later, and the current
interrogations, which cannot be
b) Dynamic session freshness:
Ensure session freshness check into
the integrity check mechanism.
c) Self/Non-self identification:
Ensure only author ized self entity can
access the network resources and
services, and eliminate any non-self
a) Authenticat i on :
Ensure that only legal entiti es can access
network resources to exclude any illegal
entity from the networks;
b) Authorization:
Realize different access control among legal
Advanced authentication and
authorization mechanisms:
a) Assignment of users contributions
to a single contributor;
b) Upgrading users trust status
without revealing identities to service
a) Intelligent access control:
Use heterogeneous authentication and
identification for semantic access
control on legal inf or mation
b) Compatible certificate authority:
Authenticate entity and grant
authority t o access sy stem resource s;
c) Hierarchical authentication:
Establish hierarchical mutual
authentication: individual/group
authentication, and source/terminal
a) Providing available proofs to prevent any entity from denying having performed a
particular behavior related to the exchanged messages;
b) Ensuring the availability of evidence that can be presented by a trusted third party,
and proving that an entity’s behavior has occurred before.
Social attribution:
Assign social attribution to an
entity’s cyber behaviors, which a re
applied for compatible social
computing and behavior supervision.
Privacy Any sensitive information is protected, that
may be derived from the observation of
network activities.
Dynamic Consent Tool:
Permit certain services or applications
to access as little or as much of that
the data as desired.
a) Transparency:
Let user know which entity contains
its related data, when and where the
entity has used the data, and how the
entity realizes the specific function.
b) Traceability:
Let the entity know the network and
service information that it has even
Copyright © 2012 SciRes. AIT
Compatibility requires that an entity has appropriate
interconnection and interoperability to adjust to hetero-
geneous data formats, interfaces, channels, and networks
in U2IoT model. The supp lemental requirements address
ad vanced criterions for information interaction. Meanwhile ,
compatibility can be promoted to scalability, expansibility
and modularity among heterogeneous entities and the mul-
ti-context environments.
The both requirements operate together to promote the
security and privacy preservation: 1) ensure diverse enti-
ties own artificial intelligence and autonomous security
control against the strong attackers; 2) ensure heteroge-
neous entities, netwo rks and applications establishing re-
liable interconnection without compromising any com-
munication data and individual privacy.
3.2. Physical Security
Physical security is denoted in external context and inh e-
rent infrastructure, in which human-lik e security immune
safeguard is achieved.
1) External Context: Simple con text and complex con-
text are specified in [16], in which the fo rmer determines
the basic identity, location, and entity status by a single pa-
rameter; the latter refers to geographical structures, tracea-
bility information, and real world cond itions. Above both
contexts are refined to support creating, debugging and
integrating applications of Ubiquitous IoT, and provide
interface interconnection and restriction for Unit IoTs. In
U2IoT model, the borders of each entity’s external con-
text merge even vanish, and the obscure contexts span-
ning from an individual, an object, or an environment to
social relationships, should support the hierarchical IoT
subsystem. Particularly, intrusion detection algorithm is
significant to acquire context information for monitoring
sensors behavior, discover control node breaches, and other
potential vuln erabilities.
2) Inherent Infrastructure: Artificial immune security
system as computational intelligence is applied to ana-
lyze inherent infrastructure, which belongs sensorial sys-
tem inspired by principles and processes of the natural
immune system. Typical algorithms (e.g., clonal selection,
negative selection, and immune network) exploit the im-
mune system’s features of detection, learning and mem-
ory to cons titute in n ate immunity an d ad ap tive immunity .
Physical security issues such as intrusion detection, adap-
tive disposition, context-driven feedback, and error reco-
very can be addressed as follows.
Innate immunity: It provides basic barriers against
foreign invasions in real-time environment, and it is
triggered upon sensors identifying abnormal or ma-
levolent attacks by the intelligent pattern recognition
mechanisms. Co-stimulation signa ls are transmit ted to
distributed control nodes via Unit IoT networks, and
then rejection reactions are performed by management
centers. During defense operations, activation thresh-
olds are defined to ensure the detection optimization,
and fuzzy diagnosis can also applied for imperfect de-
tection. Note that the innate immune defense is non-
specific, meaning that U2IoT model responds to the
various attacks in a general scheme. Such system can-
not afford long-lasting immunity against a certain at-
tack. The innate immune system is dominant to con-
front the dynamic contexts and continuously refresh-
ing threats.
Adaptive immunity: It refers to acquired resistance,
where an attack is marked as a specific signature. Se-
lective response requires recognizing non-self ele-
ment during attack prototype presentation. If U2IoT
has been infected by the same or similar invasion, spe-
cific memory module would be aroused to eliminate
damaging effects by generating improved response to
return th e system into secure state. Adap tive immunity
executes fuzzy diagnosis to variations of the same for-
mer attack, and optimal stimulation such as subsidiary
vaccination is available by updating M&DCs’ profile
According to both inna te and adaptive immun ities [17],
three main features should be achieved in U2IoT model.
Multithreaded and Hybrid Configuratio n: The U2IoT
model may apply multithreaded security algorithms for
the massively parallel network architecture that comprises
a diverse set of components. The components are organized
in hybrid mode, in which both centralized and distribu ted
configurations are included. Towards Unit IoT, the allo-
cation of the sensing and query processing is performed
by the central M&DC. Towards Ubiquitous IoT, the in-
dustry IoTs and local IoTs are relatively independent,
which commonly con stru ct natio na l IoT. In U2IoT model,
such multithreaded an d hybrid conf iguration are through -
out all the networks, sensor and control nodes, and man-
agement and data centers.
Multilayered and Autonomous Organiza tion : There is
no single security mechanism that offers complete immu-
nity. Therefore, multilayered protection should indepen-
dently operate for all-round safeguards. During the lay-
ered organization, U2IoT model autonomously makes its
decisions by detecting potential attacks and proposing fea-
sible solutions based on artificial immune algorithms.
Heterogeneity: U2IoT model should be accessible by a
large number of heterogeneous communication technolo-
gies with different networks, channels, interfaces, and hard-
ware/software capabilities. Such heterogeneity of en tities
adds complexity to its security situations, which makes
that a certain attack may simultaneously act on multiple
entities in different IoTs, but the attack cannot act on all
the involved IoTs. The immune protection ensures that the
Copyright © 2012 SciRes. AIT
entire heterogeneous components cannot be corrupted due
to the same attacker.
3.3. Management Security
Towards the future IoT, it is scarcely possible to estab-
lish a uniform security protocol as Internet, just like dif-
ferent nations and/or regions cannot adopt an identical
safety precaution. Hence, distinctive management mecha-
nisms are significant for both security and interconnect-
tion requirements. Due to the limitations of technological
approaches, appropriate management should couple with
the implementation of information security and physical
security. Security strategies working on human behaviors
should be considered to ensure that virtual cyber data is
adapted to the real physical contexts.
Application r e qu iremen t for distributed sensor and con-
trol nodes provides generic/specific protection. IPM is of
benefit to practical application security, such as historical
query, project management, risk assessment, software de-
sign, and system certification. For a specific scenario, custo-
mized requirements are assigned to describe the authorized/-
unauthorized usage in a particular organization or indi-
vidual. Additionally, application requ irement should also
be consistent with privacy preven tion which realizes that
the sensitive data is exchanged, stored, and shared with-
out revealing an y user privacy.
Industry/Local/National regulation mainly serves for
iM&DC/lM&DC/nM&DC to provide rules and guid ance
for U2IoT. It takes legal or disciplinary actions to resist
the offensive individuals or institutions which do not com-
ply with the regulations. Thereinto, industry regulation des-
cribes approaches to achieve high-level security objective
for a special industrial authority organization, such as ag-
riculture, energy, and military. For instance, in the chemi-
cal hazards medical management, the regulations require
certain parameters (e.g., temperature, vibrations, and rela-
tive proximity), caution the users for violation th resholds,
and guarantee system security by warning abnormal im-
plement and configuration. Thereinto, local regulation should
coincide with local customs and practices to adopt huma-
nistic perspectives for designing, implementing, and main-
taining the local IoTs. National regulatio n governs gu ide-
lines to realize nation-to-nation compatibility, and formal
memorandum of agreements needs to be shared across
national boundaries. Additionally, customized roles and
responsibilities can be codified among different nations.
International policy considers the global IoT consocia-
tion during conn ectivity and consistency of nM&DC and
global IoT. Moreover, international standards should be
addressed by governments to promote security confidence
and ensure interoperability. It indicates that a general in-
ternational governan ce framework with reasonable enfor-
cement policies will provide permanent mechanism towards
security protection.
4. Conclusion
In this paper, a security architecture IPM is proposed for
U2IoT model. The main purpose is to establish an inte-
grated security architecture with considerations on cyber-
physical-social world. The proposed IPM comprises three
essential security perspectives (i.e., in formation, physical,
and management), in which three-dimensional informa-
tion security model introduces social layer, and intelli-
gence and co mpatibility for security consideration; artifi-
cial immunity is applied to describe physical security; and
a series of social strategies are recommended to achieve
management security.
5. Acknowledgements
This work is jointly funded by National Natural Science
Foundation of China (NSFC) and Civil Aviation Admi-
nistration of Chin a (CAAC) (61079019), and is also sup-
ported by the Fundamental Research Funds for the Cen-
tral Universities (YWF-11-02-264).
[1] H. Ning, “RFID Major Projects and State Internet of T hi ng s
(Second Edition),” China Machine Press, Beijing, 2010.
[2] J. Ma, J. Wen, R. Huang and B. Huang, “Cyber-Individual
Meets Brain Informatics,” IEEE Intelligent Systems, Vol.
26, No. 5, 2011, pp. 30-37.
[3] L. Atzori, A. Iera and G. Morabito, “The Internet of Th i ng s :
A Survey,” Computer Network s, Vol. 54, No. 15, 2010, p p .
2787-2805. doi:10.1016/j.comnet.2010.05.010
[4] D. Bandyopadhyay and J. Sen, “Internet of Things: Ap-
plications and Challenges in Technology and Standardi-
zation,” Wireless Personal Communications, Vol. 58, No.
1, 2011, pp. 49-69. doi:10.1007/s11277-011-0288-5
[5] H. Ning, N. Ning, S. Qu, Y. Zhang and H. Yang, “Lay-
ered Structure and Management in Internet of Things,”
Proceedings of Future Generation Communication and
Networking (FGCN 2007), Jeju, 6-8 December 2007, pp.
[6] R. Roman, P. Najera and J. Lopez, “Securing the Internet
of Things,” Computer, Vol. 44, No. 9, 2011, pp. 51-58.
[7] R. Kaur, “Advances in Intrusion Detection System for
WLAN,” Advances in Internet of Things, Vol. 1, No. 3,
2011, pp. 51-54. doi:10.4236/ait.2011.13007
[8] G. P. Hancke, K. Markantonakis and K. E. May es, “Secu-
rity Challenges for User-Orien ted RFID Applications w i th i n
the ‘Internet of Things’,” Journal of Internet Technology,
Vol. 11, No. 3, 2010, pp. 307-313.
[9] R. Roman, C. Alcaraz, J. Lopez and N. Sklavos, “Key
Management Systems for Sensor Networks in the Context
of the Internet of Things,” Computers & Electrical Engi-
neering, Vol. 37, No. 2, 2011, pp. 147-159.
Copyright © 2012 SciRes. AIT
Copyright © 2012 SciRes. AIT
[10] L. D. Xu, “Information Architecture for Supply Chain
Quality Management,” International Journal of Produc-
tion Research, Vol. 49, No. 1, 2011, pp. 183-198.
[11] L. Zhou and H. C. Chao, “Multimedia Traffic Security
Architecture for the Internet of Things,” IEEE Network,
Vol. 25, No. 3, 2011, pp. 35-40.
[12] H. Ning and Z. Wang, “Future Internet of Things Archi-
tecture: Like Mankind Neural System or Social Organiza-
tion Framework?” IEEE Communications Letters, Vol. 15,
No. 4, 2011, pp. 461-463.
[13] Draft ITU-T Recommendation X.805 (Formerly X.css),
Security Architecture for Systems Providing End-to-End
communications, 2003.
[14] D. Havlik, S. Schade, Z. A. Sabeur , P. Mazzetti, K. Wat-
son, A. J. Berre and J. L. Mon, “From Sensor to Observa-
tion Web with Environmental Enablers in the Future Int er-
net,” Sensors, Vol. 11, No. 4, 2011, pp. 3874-3907.
[15] L. Atzori, A. Iera and G. Morabito, “The Internet of Thi ngs :
A Survey,” Computer Netwo rks, Vol. 54, No. 15, 2010, p p .
2787-2805. doi:10.1016/j.comnet.2010.05.010
[16] N. Koshizuka and K. Sakamura, “Ubiquitous ID: Stan-
dards for Ubiquitous Computing and the Internet of
Things,” IEEE Pervasive Computing, Vol. 9, No. 4, 2010,
pp. 98-101. doi:10.1109/MPRV.2010.87
[17] P. K. Harmer, P. D. Williams, G. H. Gunsch, G. B. Lamo nt,
“An Artificial Immune System Architecture for Computer
Security Applications,” IEEE Transactions on Evolution-
ary Computation, Vol. 6, No. 3, 2002, pp. 252-280.