Classifying Security Threats in Cloud Networking

Bruno M. Barros

, Leonardo H. Iwaya

1,2

, Marcos A. Simplicio Jr.

, Tereza C. M. B. Carvalho

Andr

as M

ehes

and Mats N

aslund

Escola Polit

ecnica, Universidade de S

ao Paulo, S

ao Paulo, Brazil

Karlstad University, Karlstad, Sweden

Ericsson Research, Stockholm, Sweden

Keywords:

Cloud Networking, Cloud Security, Security Threats, Security Taxonomy.

Abstract:

A central component of managing risks in cloud computing is to understand the nature of security threats. The

relevance of security concerns are evidenced by the efforts from both the academic community and technolog-

ical organizations such as NIST, ENISA and CSA, to investigate security threats and vulnerabilities related to

cloud systems. Provisioning secure virtual networks (SVNs) in a multi-tenant environment is a fundamental

aspect to ensure trust in public cloud systems and to encourage their adoption. However, comparing existing

SVN-oriented solutions is a difﬁcult task due to the lack of studies summarizing the main concerns of network

virtualization and providing a comprehensive list of threats those solutions should cover. To address this issue,

this paper presents a threat classiﬁcation for cloud networking, describing threat categories and attack scenar-

ios that should be taken into account when designing, comparing, or categorizing solutions. The classiﬁcation

is based on the CSA threat report, building upon studies and surveys from the specialized literature to extend

the CSA list of threats and to allow a more detailed analysis of cloud network virtualization issues.

1 INTRODUCTION

The current concept of cloud computing evolved

from technologies such as distributed computing and

resource virtualization, enabling the utilization of

shared computing infrastructures for delivering soft-

ware, platforms and infrastructures to different cus-

tomers over the Internet. Nevertheless, cloud com-

puting has other particular requirements such as (Mell

and Grance, 2011): on-demand provision of the com-

puting resources; broad network access to conﬁg-

ure and request computing capabilities; resources are

pooled to be used by multiple customers in a multi-

tenant model; the resources should be elastically pro-

visioned and released; and delivered services should

be transparently measured for managing and billing

purposes. This new model of delivering computing

power takes advantage of economies of scale, allow-

ing cloud providers to deliver services for a reason-

able cost to several institutions and companies. It also

brings advantages to customers, who can pay only for

what they consume instead of obliging them to pur-

chase, install and maintain their own equipment.

Unfortunately, however, the advantages brought

by the cloud are also accompanied by threats and se-

curity vulnerabilities that discourage its full adoption

by many companies.An example is the need of iso-

lating resources, data and communication within the

cloud. Public cloud systems utilize a multi-tenant ar-

chitecture, in which customers should only ”see” the

cloud resources assigned to them, as if they were the

sole user of the infrastructure.

Virtualization technologies play a crucial role in

enforcing this isolation, given that they are the main

building block in provisioning the customers’ infras-

tructure, including virtual machines (VMs) and vir-

tual networks (VNs). Additionally, a virtualization

solution(s) should ensure not only that the VMs op-

erate with isolated resources, but also allow network

trafﬁc monitoring and the creation of secure network

domains. For this reason, enabling SVN in the cloud

computing is currently a subject of intense research

(Sun and Hu, 2012). Many of the existing propos-

als rely on open network virtualization solutions such

as Open vSwitch for deﬁning virtualized network ar-

chitectures with security features (Hao et al., 2010;

Cohen et al., 2013), inserting security modules in-

side VMs and virtual switches (Basak et al., 2010;

214

M. Barros B., H. Iwaya L., A. Simplício Jr. M., C. M. B. Carvalho T., Méhes A. and Näslund M..

Classifying Security Threats in Cloud Networking.

DOI: 10.5220/0005489402140220

In Proceedings of the 5th International Conference on Cloud Computing and Services Science (CLOSER-2015), pages 214-220

ISBN: 978-989-758-104-5

 2015 SCITEPRESS (Science and Technology Publications, Lda.)

Barjatiya and Saripalli, 2012), or creating hypervisor-

based network controllers (Mattos and Duarte, 2013).

Nonetheless, it is often hard to clearly identify all

the threats and vulnerabilities that are addressed by

the different network virtualization solutions. Indeed,

there is a myriad of issues that can be targeted, such

as ensuring trafﬁc isolation, preventing snifﬁng and

address spooﬁng, as well as detecting and mitigating

Distributed-/Denial-of-Service DoS/DDoS and man-

in-the-middle attacks (Hao et al., 2010; Basak et al.,

2010; Barjatiya and Saripalli, 2012; Cohen et al.,

2013; Mattos and Duarte, 2013). However, it is not

always the case that solutions proposed in the liter-

ature explicitly analyze their (in)ability to cope with

each of the existing threats, even if they are truly able

to prevent them. This makes comparing and evaluat-

ing these solutions a difﬁcult task.

Aiming to address this lack of uniformity in the

treatment of network virtualization security propos-

als, this paper presents a threat classiﬁcation for

SVNs, describing threat classes and attack scenar-

ios that should be taken into account when design-

ing, comparing, categorizing or evaluating solutions.

This classiﬁcation is based on technical reports from

cloud standardization organizations such as European

Network and Information Security Agency (ENISA,

2013), the Cloud Security Alliance (CSA, 2013),

and National Institute of Standards and Technol-

ogy (NIST, 2011), as well as on scientiﬁc papers

that reviewed problems in this area (Chowdhury and

Boutaba, 2010; Pearce et al., 2013). Given the broad

scope of the security challenges described in these re-

ports, they do not (intend to) provide a framework

to evaluate security issues directly related to network

virtualization in cloud computing. Therefore, the

classiﬁcation proposed herein aims to ﬁll part of this

gap by focusing speciﬁcally on the technical issues re-

lated to virtual networking in the cloud environment.

The rest of this paper is organized as follows.

Section 2 reviews the security threat classiﬁcation

for cloud computing proposed by CSA (CSA, 2013).

Section 3 presents our proposed threat classiﬁcation,

which builds upon the CSA work. Section 4 then dis-

cusses, by means of examples, different cloud virtual

networking attacks from each threat category, show-

ing the coverage of the proposed classiﬁcation. Sec-

tion 5 presents the conclusion and future work.

2 CLOUD SECURITY THREATS

With the widespread adoption and popularization of

cloud-based systems, considerable effort has been

made to identify and classify security threats in this

environment. Some relevant examples include the

security guidelines for cloud computing provided by

the ENISA (Catteddu, 2010; ENISA, 2013), the CSA

(CSA, 2013), and NIST (NIST, 2011).

Among these documents, the CSA “Notorious

Nine” report (CSA, 2013), which identiﬁes important

threats that may occur accidentally or intentionally in

cloud systems, is of especial interest: it provides a

clear view of the most relevant security threats when

deploying and consuming cloud services, ranked ac-

cording to the industry perspective. Therefore, the re-

port highlights the main security aspects that need to

be taken into account by cloud providers for ensur-

ing trust in their services. It is important to notice,

however, that (CSA, 2013) is intended as a general

guideline of relevant security aspects, not focused on

networking issues. Nonetheless, given its importance,

it can be seen as an interesting starting point for iden-

tifying relevant cloud networking threats, which is ex-

actly the approach adopted in this document. Next,

we present a classiﬁcation method for indentify secu-

rity threats in cloud networking, built upon the CSA

classiﬁcation for cloud security threats.

3 THREAT CLASSIFICATION

CSA “Notorious Nine” (CSA, 2013) and similar-

purpose reports (CSA, 2011; ENISA, 2013; NIST,

2011; Gonzalez et al., 2012) are important sources

of information about cloud security threats. However,

their main goal is to give a high level description of

potential problems, not on providing a ﬁne-grained

analysis of how each of the many threats identiﬁed

apply to speciﬁc scenarios (e.g., virtual networking).

On the other hand, there are works in the literature

that investigate SVNs in more depth, such as (Chowd-

hury and Boutaba, 2010; Schoo et al., 2011; Natara-

jan and Wolf, 2012). Unfortunately, since their goal

is to survey solutions and to identify challenges in the

area, they fail to provide a reusable classiﬁcation of

the virtual networking threats described. Given the

relevance of threat modeling to identify security re-

quirements (Myagmar et al., 2005), those works are

not ideal for the task of comparing and evaluating

different security proposals in virtual networking or

guiding the design of comprehensive solutions for the

most relevant threats. Aiming to bridge this gap, next,

we build upon the CSA “Notorious Nine” threat re-

port for deriving ﬁner-grained threat classiﬁcation fo-

cused speciﬁcally on SVNs.

ClassifyingSecurityThreatsinCloudNetworking

215

3.1 Extending Current Classiﬁcations

For the purpose of building the proposed classiﬁ-

cation, the CSA “Notorious Nine” threats were de-

composed into more speciﬁc menaces. The result-

ing ﬁner-grained list, containing speciﬁc attacks and

countermeasures, was then analyzed aiming to iden-

tify threats that could be associated to virtual network-

ing security issues. The aggregation of the identiﬁed

threats according to their characteristics then lead to

the general categories presented as follows.

Before we present the proposed classiﬁcation,

however, it is important to emphasize that attaining

the right level of abstraction is a considerable chal-

lenge when trying to create a comprehensive view of

virtual networking vulnerabilities in the cloud. Aim-

ing to be both concise and comprehensive, our ap-

proach in the proposed classiﬁcation involves two

main requirements: (1) threats should have a detailed

enough description to effectively help guide the de-

velopment of innovative solutions; and (2) the number

of threat groups should be small enough to allow the

classiﬁcation to be applied to the analysis and com-

parison of common solutions. These requirements

have an obvious trade-off: a large number of threat

classes may lead to an overly detailed classiﬁcation,

but a reduced number of threat groups may lead to a

high level description that may be vague and less use-

ful to comparative studies. As a result, our classiﬁca-

tion proposes a reduced number of categories without

ignoring important aspects of virtual networking.

In addition to those basic requirements, we con-

sidered the different attack scenarios within the cloud

environment, i.e., who the attacker is and who is being

attacked. In each scenario, we then try to identify the

different threat classes to reﬂect the concerns already

evidenced in the literature.

3.2 Threat Scenarios

The ﬁrst is the Cloud Provider Network, which in-

cludes all the cloud provider private network re-

sources connecting all the data center infrastructure

that allows the cloud service provision. The second is

the Public Network, which comprises the public In-

ternet that allows users to access the cloud services.

Both networks are illustrated in Figure 1. We have

limit our scope to threats in the Cloud Provider Net-

work, to identify security issues related only to the

cloud computing paradigm and to its technological

mechanisms. We can identify three attack scenarios

involving different entities of the cloud:

1. Tenant-to-Tenant. Threats related to attacks pro-

moted by a legitimate tenant targeting another le-

gitimate tenant. Such attacks are usually per-

formed by exploring vulnerabilities of the cloud

provider network infrastructure

2. Tenant-to-Provider. Threats related to cloud vul-

nerabilities that allow a legitimate tenant to dis-

rupt the operation of the cloud infrastructure, pre-

venting the cloud provider from delivering the ser-

vice in accordance with the service level agree-

ments established with other legitimate tenants.

3. Provider-to-Tenant. Threats related to vulnera-

bilities in the cloud provider infrastructure, which

allows malicious insider attacks from employees

and other agents with direct access to the cloud

infrastructure.

Figure 2 illustrates these three attack scenarios in

the network context previously mentioned. Despite

the interest of provider-to-tenant threats, for the pur-

pose of this research we consider a reliable cloud

provider, focusing our efforts on the tenant-to-tenant

and tenant-to-provider threats.

3.3 Cloud VN Threat Categories

Following the method and the requirements discussed

in Section 3.1, we identiﬁed ﬁve general classes of

virtual networking security threats, described as fol-

lows. For each class, we also present the CSA-related

threats and the virtual networking attacks obtained

from the decomposing process described in Section

3.1. We note, however, that the inclusion of new

threat classes to this classiﬁcation should be consid-

ered a matter of continuous work, following new dis-

coveries in the ﬁeld of cloud security.

1. Physical Isolation. Covers all the vulnerabilities

related to the physical resources of the underlying

network infrastructure being shared by multiple

tenants. Attacks in this class are normally related

to capturing and to analyzing data collected from

shared resources, but can also involve the exhaus-

tion of resources from shared hardware.

2. Logical Isolation. Covers all the vulnerabilities

directly related to situations in which logical re-

sources (e.g., vCPUs, vLANs and vSwitches) are

not adequately isolated, allowing tenants to access

each other’s networking capabilities. Attacks of

this class can exploit vulnerabilities in the cloud

Figure 1: Complementary networks in cloud environments.

CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience

216

Figure 2: Security attack scenarios in cloud VNs.

virtualized network functions and network man-

agement modules.

3. Authentication. Covers all the vulnerabilities re-

lated to inadequate authentication, which allow at-

tackers to mask their real identities. This can be

accomplished by exploiting authentication proto-

cols, acquiring credentials and/or key materials by

capturing data trafﬁc, or via password recovery at-

tacks (e.g., brute force or dictionary attacks).

4. Authorization. Covers all the vulnerabilities re-

lated to authorization problems, allowing granting

or scaling rights, permissions or credentials to or

from an unauthorized user. The attacker can ex-

ploit a vulnerability in the cloud platform autho-

rization modules, or even in the victim computer,

to create or change its credentials in order to ob-

tain privileged rights. Similarly to attacks that ex-

ploit authentication threats, authorization threats

can lead to the leakage of conﬁdential data and

access to management modules.

5. Insecure APIs. Covers all the threats related to

failures, malfunctions and vulnerabilities in APIs

that compose the cloud system. Attacks of this

class try to exploit insecure interfaces for access-

ing or tampering with services running in other

tenants or cloud administrative tools. This may

lead to data loss/leakage, as well as to cause un-

availability of services.

3.4 Correlation with CSA Threats

Table 1 summarizes the correlations between CSA

threats and the network threat categories herein pro-

posed. Analyzing this table, one can notice that

CSA threats related to shared technology vulnera-

bilities, data breaches, and DoS are the most fre-

quently related to network vulnerabilities. Therefore,

they should be the main focus of SVN-oriented solu-

tions, considering both tenant-to-tenant and tenant-to

provider attack scenarios, as discussed in more details

in the next section.

Table 1: Correlation of proposed categories and threats with

CSA notorious top nine threats.

Category Threat Examples CSA Top 9

Side-channel attack DB,

Physical Guest-hoping attack DoS,

Isolation Resource exhaustion STV

Physical man-in-the-middle

Man-in-the-middle DB,

Logical Snifﬁng DoS,

Isolation Port scanning STV

Replay attacks

Authentication

Spooﬁng DB,

Compromised-key attack ASH,

Password-based attacks STV

Authorization

Software exploitation DB, MI,

Phishing ACS, STV

Insecure APIs

API exploitation

IIA

Code and Package injection

Abbreviations: data breaches (DB), Denial-of-Service (DoS),

shared technology vulnerabilities (STV), account or service

trafﬁc hijacking (ASH), malicious insiders (MI), abuse of

cloud services (ACS), insecure interfaces and APIs (IIA).

4 CLASSIFICATION ANALYSIS

This section correlates the threat scenarios discussed

in Section 3.2 with the classiﬁcation proposed in Sec-

tion 3.3, presenting threat examples for each pro-

posed category in both Tenant-to-Tenant and Tenant-

to-Provider scenarios.

4.1 Physical Isolation

Such attacks can be perpetrated if the attacker’s VM

is sharing the same host machine or physical network

node as the victim’s VM. In this scenario, the attacker

can engage in: a side-channel attack, subjecting conﬁ-

dential trafﬁc to cryptanalysis; use DoS/DDoS attacks

to exhaust or to bring physical network resources

down; or to gather information about the services run-

ning in the target machine. Snifﬁng is one of the most

common attacks: attackers attempt to access informa-

tion from other tenants or cloud provider network by

reading packets going through the physical NICs.

Threat Examples on Tenant-to-Tenant Scenario.

1) Side channel: An encryption algorithm used by

ClassifyingSecurityThreatsinCloudNetworking

217

tenants is subject to cryptanalysis based on timing at-

tacks. 2) Guest hopping: The attacker inserts a VM in

the same host of the victim to exploit shared network

components in subsequent attacks. 3) DoS/DDoS: Ex-

haustion of network resources shared by tenants in-

side the same host.

Threat Examples on Tenant-to-Provider Scenario.

1) Snifﬁng: By monitoring physical network inter-

faces an attacker may intercept provider management

data. 2) Side channel: An encryption algorithm used

by a provider is subject to cryptanalysis based on tim-

ing attacks. 3) DoS: Ampliﬁcation attacks performed

from inside the cloud, possibly combined with ad-

dress spooﬁng can ﬂood provider network infrastruc-

ture.

4.2 Logical Isolation

The cloud provider should ensure isolation of re-

sources among its customers, which includes virtual-

ized network components that compose the cloud log-

ical network and the tenants virtual networks. Tenant-

to-tenant attacks involves breaking this logical isola-

tion, allowing a malicious tenant to get access vic-

tim’s resources. This sort of attack can lead to data

leakage, malicious use of other tenants’ resources,

and/or disruption of network services.

A malicious tenant may also exploit virtual net-

work resources, controllers and other services pro-

vided by the cloud provider. Some examples involve

the use of port scanning and network reconnaissance

mechanisms, so that the attacker can get access to

privileged information, such as the network topology,

list of virtual networks, or conﬁguration messages.

Also, a malicious tenant may use the cloud infras-

tructure to deploy botnets for launching DoS attacks

against the cloud infrastructure, its tenants or any ex-

ternal targets.

Threat Examples on Tenant-to-Tenant Scenario.

1) Port scanning: The attackers scans open ports and

running services in other tenant VMs. 2) Network

reconnaissance: Protocols for network conﬁguration

and identiﬁcation can be exploited to discover other

virtual networks and/or network topologies pertaining

to other tenants (e.g., using ARP requests) 3) Snifﬁng:

Snifﬁng and/or man-in-the-middle attacks due to the

lack of isolation between virtual networks. 4) Mal-

ware (worm): Malicious software that can replicate

itself through and between tenant virtual networks.

5) Botnets: Use the cloud infrastructure to deploy bot-

nets, in which groups of VMs running malicious com-

puter programs target other tenants and shared net-

work resources.

Threat Examples on Tenant-to-Provider Scenario.

1) Network reconnaissance: Use network reconnais-

sance mechanisms to discover other network topolo-

gies, VNs, and services. 2) Malware (worm): Ma-

licious software that can replicate itself through the

provider network infrastructure from a compromised

tenant virtual network. 3) Botnets: Use the cloud in-

frastructure to deploy botnets, targeting shared net-

work resources and network controller nodes. 4) Re-

play attack: replication of control messages that can

impair network services and cloud operations (e.g.,

instantiating duplicated VMs)

4.3 Authentication

Attacks that exploit authentication vulnerabilities can

be performed either in Tenant-to-Tenant and Tenant-

to-Provider scenarios. In both cases the attacker

may exploit the authentication protocols by means of

compromised-key (e.g., leakage of key material infor-

mation) and/or password-based attacks (e.g., dictio-

nary and brute force attacks). Also, the authentication

and identity management modules can be exploited to

force authentication threats. In the case of Tenant-to-

Tenant scenario, the attacker wants to gain access to

the victim’s virtual networks. In Tenant-to-Provider

scenario, the attacker may gain access to the whole

network services and/or controlling modules, affect-

ing the whole cloud.

Threat Examples on Tenant-to-Tenant Scenario.

1) Credential replay: By capturing and replaying a

user’s credential. 2) Spooﬁng: The malicious tenant

masquerades as another tenant by falsifying data and

thereby gaining illegitimate access to resources.

Threat Examples on Tenant-to-Provider Scenario.

1) Credential replay: The replay of a user’s credential

might allow impersonation of a cloud administrator.

2) Spooﬁng: The malicious tenant performs a DNS

spooﬁng attack, inserting bogus data into a DNS name

server cache database and causing the name server to

return an incorrect IP address.

4.4 Authorization

In this case, the attackers try to escalate their privi-

leges in the system. Since network privileges in the

cloud usually make sense only with respect to the

cloud provider, authorization attacks are reasonable

only when we consider Tenant-to-Provider scenario.

For instance, a malicious user may fake its creden-

tials to acquire administrative roles, authorization to

other customers services, or to gain access to services

out of contract.

Threat Examples on Tenant-to-Provider Scenario.

1) Network Reconnaissance: A legitimate tenant can

CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience

218

escalate its privileges to run network reconnaissance

scripts in the cloud network infrastructure, discover-

ing the cloud provider network topology and using

this information for promoting more precise attacks

against the network resources. 2) Cloud Exploit Kits

(Malware-as-a-Service): Malicious software hosted

inside cloud provider infrastructure and available to

other tenants to attack the provider services through

its network resources. 3) Network Programmability:

The attackers try to escalate their privileges to get ac-

cess to program APIs of the network devices (e.g.,

OpenFlow API).

4.5 Insecure APIs

Attacks to APIs can affect a broad range of cloud

modules, e.g., those responsible for resource alloca-

tion, authentication and identity management, stor-

age, or accounting. The network modules deployed

in VMs, as well as controller, computing and network

nodes, are usually distributed along the cloud infras-

tructure. Therefore, the APIs used for network vir-

tualization and conﬁguration may be target of attacks

and vulnerability exploitation. For instance, attacks

based on code injection techniques may exploit com-

puter errors caused by processing invalid data, under-

mining cloud services and databases. This category

of attack may target either Providers or Tenants.

Threat Examples on Tenant-to-Tenant Scenario:

1) Code injection: The attacker performs SQL injec-

tion using a network controller API (e.g., Neutron) to

erase a tenant data from the cloud network conﬁgura-

tion database.

Threat Examples on Tenant-to-Provider Scenario:

1) Code injection: The attacker performs SQL injec-

tion using a network controller API (e.g., Neutron) to

modify (parts of) the network conﬁguration database.

5 CONCLUSIONS AND FUTURE

WORK

The wide variety of threats related to cloud comput-

ing network virtualization makes it difﬁcult to com-

pare or to categorize existing solutions focused on se-

curing virtual networks. This paper proposes a threat

classiﬁcation for cloud virtual networks built upon

the “notorious nine cloud computing top threats” of

CSA. Moreover, the presented classiﬁcation allows a

more detailed view of the network threats discussed

in cloud computing literature. This ﬁner-grained ap-

proach makes it easier to identify the technologies

that might be used to solve different security issues

in cloud networking, facilitating the analysis and de-

sign of security solutions.

As future work we plan to employ this threat clas-

siﬁcation in a literature review of cloud networking

security solutions. The result expected is a compre-

hensive literature survey that allows not only com-

paring existing solutions, but also the identifying the

gaps and challenges in cloud networking security.

ACKNOWLEDGEMENTS

Innovation Center, Ericsson Telecomunicac¸

oes S.A.

(Brazil) and CNPq (grant 305350/2013-7).

REFERENCES

Barjatiya, S. and Saripalli, P. (2012). BlueShield: A Layer 2

Appliance for Enhanced Isolation and Security Hard-

ening among Multi-tenant Cloud Workloads. IEEE

Int. Conf. on Utility and Cloud Comp., pages 195–198.

Basak, D., Toshniwal, R., Maskalik, S., and Sequeira, A.

(2010). Virtualizing networking and security in the

cloud. SIGOPS Oper. Syst. Rev., 44(4):86–94.

Catteddu, D. (2010). Cloud computing: Beneﬁts, risks and

recommendations for information security. In Serr

ao,

C., Aguilera D

ıaz, V., and Cerullo, F., editors, Web

Application Security, volume 72 of CCIS, page 17.

Chowdhury, N. and Boutaba, R. (2010). A survey of net-

work virtualization. Comput. Netw., 54(5):862–876.

Cohen, R., Barabash, K., Rochwerger, B., Schour, L.,

Crisan, D., Birke, R., Minkenberg, C., Gusat, M., Re-

cio, R., and Jain, V. (2013). An intent-based approach

for network virtualization. In IFIP/IEEE INM’13.

CSA (2011). Security Guidance for Critical Areas of Focus

in Cloud Computing V3.0. Technical report, CSA.

CSA (2013). The Notorious Nine Cloud Computing Top

Threats in 2013. Technical report, CSA.

ENISA (2013). Threat landscape 2013-overview of current

and emerging cyber-threats. Technical report, ENISA.

Gonzalez, N., Miers, C., Red

ıgolo, F., Jr. Simplicio, M.,

Carvalho, T., N

aslund, M., and Pourzandi, M. (2012).

A quantitative analysis of current security concerns

and solutions for cloud computing. JCC, 1(1):1–18.

Hao, F., Lakshman, T. V., Mukherjee, S., and Song, H.

(2010). Secure Cloud Computing with a Virtualized

Network Infrastructure. In Proc. of the USENIX.

Mattos, L. F. D. and Duarte, O. C. M. B. (2013). A Mech-

anism for Secure Virtual Network Isolation Using to

Hybrid Approach Xen and OpenFlow. In SBSeg’2013.

Mell, P. and Grance, T. (2011). The NIST deﬁnition of

cloud computing (draft). Technical report, NIST.

Myagmar, S., Lee, A., and Yurcik, W. (2005). Threat mod-

eling as a basis for security requirements. In SREIS.

Natarajan, S. and Wolf, T. (2012). Security issues in net-

work virtualization for the future internet. In ICNC.

ClassifyingSecurityThreatsinCloudNetworking

219

NIST (2011). Guide to Security for Full Virtualization

Technologies. Technical report, NIST.

Pearce, M., Zeadally, S., and Hunt, R. (2013). Virtual-

ization: Issues, security threats, and solutions. ACM

Computing Surveys (CSUR), 45(2):17.

Schoo, P., Fusenig, V., Souza, V., Melo, M., Murray, P.,

Debar, H., Medhioub, H., and Zeghlache, D. (2011).

Challenges for cloud networking security. In MNM.

Sun, Q. and Hu, Z. (2012). Security for networks virtual

access of cloud computing. In MINES’2012.

CLOSER2015-5thInternationalConferenceonCloudComputingandServicesScience

220