Providing Security SLA in Next Generation Data Centers with SPECS:

The EMC Case Study

Valentina Casola

, Massimiliano Rak

, Silvio La Porta

and Andrew Byrne

Universit

a “Federico II” di Napoli, Dipartimento di Ingegneria Elettrica e Tecnologie dell’Informazione, Napoli, Italy

Seconda Universit

a di Napoli, Dipartimento di Ingegneria dell’Informazione, Aversa, Italy

EMC Ireland COE Innovation, Cork, Ireland

Keywords:

Cloud, ngDC, Cloud Security, Security SLA.

Abstract:

Next generation Data Centers (ngDC) are the cloud-based architectures devoted to offering infrastructure

services in ﬂexible ways: managing in an integrated way compute, network and storage services. This solution

is very attractive from an organisation’s perspective but one of the main challenges to adoption is the perception

of loss of security and control over resources that are dynamically acquired in the cloud and that reside on

remote providers. For a full adoption, datacenter customers need more guarantees about the security levels

provided, creating the need for tools to dynamically negotiate and monitor the security requirements. The

SPECS project proposes a platform that offers security features with an as-a-service approach, furthermore it

uses Security Service Level Agreements (Security SLA) as a means for establishing a clear statement between

customers and providers to deﬁne a mutual agreement. This paper presents an industrial experience from

EMC that integrates the SPECS Platform and their innovative solutions for ngDC. In particular, the paper will

illustrate how it is possible to negotiate, enforce and monitor a Security SLA in a cloud infrastructure offering.

1 INTRODUCTION

Storage services, like many IT services, are in-

creasingly moving toward the virtualized, distributed

cloud model. Indeed, recent concepts like ngDC or

Software-Deﬁned Data Centers (SDDC) (Davidson,

2013) are grounded in the ideas of virtualization, of-

fering the capability to run multiple independent vir-

tual servers using a set of shared, physical resources.

Resource Pooling is another signiﬁcant advantage of

the ngDC and SDDC solutions, enabling the auto-

matic allocation of storage, network and compute re-

sources to meet the demand of incoming requests.

This is one of the foundational concepts cloud com-

puting is built on. In fact, through ngDC provisioning

models, a Cloud Service Provider (CSP) may offer

on demand, scalable, secure and cost effective cloud

infrastructures or services upon which Cloud Service

Customers (CSC) can develop their own services.

However, one of the key limiting factors holding

back larger adoption of the cloud services is trust. In

cloud computing the tangible assets of the CSC be-

come intangible, virtual resources, that are dynami-

cally acquired via a 3

Party provider. With these

provisioning models, the loss of control over their

own services and assets (data), CSCs are naturally

hesitant to place critical business applications or sen-

sitive data in the cloud. A possible solution to this

challenge could be the adoption of SLAs, clearly stat-

ing what services are provided by the CSP and the

related responsibilities in case of violation.

Despite the intense research efforts into develop-

ing standards and frameworks for SLAs (Theilmann

et al., 2008), (Morin, 2011),(EC, 2011), (CSCC,

2012), (International Organization for Standardiza-

tion, 2014), at the state of the art few solutions al-

low CSPs to offer practical, implementable Security

SLAs. Moreover, there are very few services able to

concretely monitor the security features which would

enable CSCs to verify the status of guaranteed SLAs.

In such a context, the SPECS project

proposes

a framework which aims to facilitate the automated

negotiation, monitoring and enforcement of Security

SLAs. This paper illustrates how the SPECS frame-

work is used in order to enhance a ngDC storage ser-

vice with security features based on customer’s re-

quirements. In particular, we integrate SPECS with

http://www.specs-project.eu/

138

Casola, V., Rak, M., Porta, S. and Byrne, A.

Providing Security SLA in Next Generation Data Centers with SPECS: The EMC Case Study.

In Proceedings of the 6th International Conference on Cloud Computing and Services Science (CLOSER 2016) - Volume 2, pages 138-145

ISBN: 978-989-758-182-3

the ViPR storage controller

, a commercial product

offered as a service by EMC, in order to fully imple-

ment the ngDC paradigm by offering storage services

protected by Security SLAs.

The implementation of security capabilities, en-

forced and monitored through SPECS demonstrates

the effectiveness of Security SLAs as a concrete so-

lution that can be adopted in commercial products.

The effectiveness and adaptability of this approach is

further strengthened by use of standardized security

metrics to identify capabilities delivered by SPECS to

support the storage service. The outcome of this pro-

cess is to improve trust in the capabilities of the CSP

to protect and guarantee their assets services in the

cloud.

The remainder of this paper is organized as fol-

lows: Section 2 introduces the SPECS framework for

the provisioning of cloud services guaranteed by Se-

curity SLAs. Section 3 describes the main features

and limitations of ngDC that motivate the need for

more ﬂexible and secure Data Centers. Section 4 in-

troduces our proposal of enhancing the ngDC with

Security SLAs based on the adoption of SPECS. In

particular, this section focuses on EMC storage so-

lutions. Section 5 describes the architecture of the

ngDC storage testbed and the enhanced security fea-

tures. Finally, Section 6 gives an overview of related

work on frameworks and guidelines for SLAs, while

Section 7 summarizes the conclusions and provides

direction for future work.

2 THE SPECS FRAMEWORK

The SPECS framework provides services and tools to

build applications offering services with security fea-

tures deﬁned in, and granted by a Security SLA (Ca-

sola et al., 2014),(Rak et al., 2013).

The framework addresses both CSPs’ and users’

needs by providing tools for 1) enabling user-centric

negotiation of security parameters in a Security SLA;

2) providing a trade-off evaluation process among

CSPs; 3) real time monitoring of the fulﬁlment of

SLAs agreed with CSPs; 4) notifying both End-users

and CSPs in the event that an SLA is violated; 5) en-

forcing agreed SLAs in order to maintain the agreed

security levels. The SPECS framework is also able

to ”react and adapt” in real-time to ﬂuctuations in the

security level by applying the required countermea-

sures.

In order to provide security capabilities granted by

Security SLAs, a SPECS Application orchestrates the

http://www.emc.com/vipr

so called SPECS Core Services dedicated to the Ne-

gotiation, Enforcement and Monitoring of an SLA.

Through these core services, the cloud service is en-

hanced with security capabilities guaranteed by the

signed SLA

In SPECS, four primary actors have been deﬁned:

• End-user: The CSC of a Cloud service;

• SPECS Owner: The Cloud service provider;

• External CSP: An independent (typically public)

CSP, unaware of the SLA, providing only basic

resources without security guarantees;

• Developer: Supports the SPECS Owner in the de-

velopment of SPECS applications.

Figure 1: SPECS Entity Relationships.

As illustrated in Figure 1, the interactions among

the parties are very simple: the End-user uses the

Cloud services offered by the SPECS Owner, which

acquires resources from External CSPs, enriched with

capabilities to meet the End-user’s security require-

ments. SPECS then monitors and enforces the End-

user’s security requirements to ensure the agreed se-

curity levels and alert the End-user of any breaches in

the terms of the Security SLA.

3 NEXT GENERATION DATA

CENTER STORAGE

The ngDC is a highly efﬁcient and optimized data

center that allows organisations to achieve more

within the conﬁnes of the available resources (phys-

ical servers, power, cooling, facilities, etc.). The key

advantage of the ngDC is its agility and ability to

adapt rapidly to changes in an organizations business

and workload requirements.

This efﬁciency is achieved in a ngDC by consoli-

dating the physical resources, in other words virtual-

izing it. As illustrated in Figure 2, we view the clas-

sical data center model as one in which there are ded-

icated physical resources for each application. The

Providing Security SLA in Next Generation Data Centers with SPECS: The EMC Case Study

139

Figure 2: Evolution of the data center towards a fully Virtualized environment.

ﬁrst step in evolving towards the ngDC is to move

away from dedicated resources to consolidated re-

sources virtualizing the physical compute resources

though the use of Hypervisor technologies.

The second evolutionary step consists of virtualiz-

ing the network resources, dividing them into discrete

segments to isolate and segregate the trafﬁc and/or the

service by creating virtual networking components

(e.g. virtual LANs, virtual SANs, virtual switches,

etc.) that are part of a Hypervisor, logical links, and

even converged networks.

The ﬁnal phase before achieving a completely vir-

tualized data center is to abstract out the physical stor-

age resources. In the classic storage model, an Intel-

ligent Storage System is used to group disks together

and then partition those physical disks into discrete

logical disks. These logical disks are assigned a Log-

ical Unit Number (LUN), and are presented to a host,

or hosts, as a physical device. Redundancy of the data

stored on the disks is provided by RAID (Redundant

Array of Independent Disks) technology, which is ap-

plied at either the physical disk layer or the logical

disk layer.

This classic model has several limitations how-

ever. For example, there is an upper limit to the max-

imum number of physical disks that can be combined

to form a logical disk. Another issue is that often the

amount of storage provisioned for each application is

greater than what is actually needed in order to pre-

vent application downtime. Both of these situations

results in inefﬁcient usage of the physical storage re-

sources that needlessly remain idle.

These kinds of inefﬁciencies can be resolved by

introducing Software Deﬁned Storage (SDS) appli-

cations such as EMC’s ViPR in order to virtualize

the tiered storage resources. The outcome is a more

efﬁcient usage of resources which result in reduced

power and space costs in the data center as well as

reduced workloads for storage and server administra-

tors. The power of these SDS solutions is the abstrac-

tion of the physical resources by creating resource

pools designed to support more generalized work-

loads across applications. This enables the capacity

usage and requirements to be more closely monitored

and aligned to the available resources - further reduc-

ing the operational costs.

But what about software-deﬁned security? Most

(or all) of the security controls can be automated and

managed through software, depending on how vir-

tualized the infrastructure is. Such an approach re-

quires any service to be controlled under some secu-

rity policy. The innovative idea proposed in SPECS

to enhance the ngDC, is to provide Security-as-a-

Service (SecaaS) according to agreed Security SLA.

To achieve this, in following sections, the integration

of the ngDC with the SPECS framework is proposed

such that the full Security SLA life cycle can be man-

aged.

4 INTEGRATING SPECS WITH

ngDC

In a ngDC, the infrastructure is virtualised, delivered

as a service and controlled by management applica-

tions. This shift in the architectural approach for the

data center offers a more agile, ﬂexible and scalable

model. The core idea is the decoupling of the hard-

ware from the software layer. Indeed, services (OSs,

applications and workloads) view these abstracted,

virtual resources as though they were physical com-

pute, storage and network resources. Figure 3 il-

lustrates the relationships between the infrastructure

layer, including different physical resources, and the

service and application layers.

Despite the multitude of beneﬁts available when

leveraging a Cloud infrastructure, wide scale Cloud

adoption for sensitive or critical business applications

CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science

140

Figure 3: Next Generation Data Center Architecture.

still faces resistance due to concerns over the pri-

vacy and security of the data, workloads and appli-

cations outsourced to the Cloud provider’s data cen-

ters. Cloud providers typically offer a set of security

measures advertised to potential customers, however

without the ability to provide assurance of those se-

curity measures or to maintain visibility over the ser-

vice, the Cloud customer has no way to verify the ser-

vice is being provided as described.

Integrating SPECS with the ngDC offers a Se-

caaS solution, not only by establishing a process to

negotiate and monitor services running in the data

center, but also building on the native security fea-

tures present by providing additional security features

delivered through virtual machines dynamically allo-

cated and instantiated on the data center to meet the

security requirements. Combined with the guarantees

provided through a Security SLA, these security fea-

tures improve the conﬁdence with which end users

can migrate their applications and data to the Cloud.

Figure 4 illustrates where the SPECS platform in-

tegrates with a typical ngDC architecture. In EMC’s

use case, ViPR is used to offer software deﬁned stor-

age management for the ngDC storage resources.

Figure 4: SPECS Enhanced Data Center Architecture.

4.1 Providing ngDC using SPECS Core

Services

In Figure 5 the usage of SPECS to provide ngDC

services is illustrated: the SPECS platform is used

to negotiate an SLA for storage resources provided

by ViPR. On receipt of the SPECS negotiated offer,

the End-user can sign the SLA which will trigger the

Enforcement phase of SPECS, making the newly ac-

quired resources available to the End-user. In parallel

to this, the Monitoring module is conﬁgured so that

the resources are continuously monitored. The three

phases are described in detail in the following subsec-

tions.

Figure 5: SPECS Service Negotiation/Brokering Service.

4.1.1 Negotiation

Formatted according to the SPECS Security SLA for-

mat (De Benedictis et al., 2015), security features are

represented using few simple concepts: Security ca-

pabilities, the set of security controls (NIST, 2013)

that a security mechanism is able to enforce on the

target service; Security metrics, the standard of mea-

surement adopted to evaluate security levels of the

services offered; Security Level Objectives (SLOs),

the conditions, expressed over security metrics, rep-

resenting the security levels that must be respected

according to the SLA.

Security-related SLOs are negotiated based on the

SPECS Customer’s requirements. A set of compli-

ant offers, each representing a different supply chain

to implement, is identiﬁed and validated. The agreed

terms are included in a Security SLA that is signed by

the SPECS Customer and the SPECS Owner.

4.1.2 Enforcement

Once the SLA has been successfully negotiated, it

is implemented through the Enforcement services,

which acquire resources from External CSPs and ac-

tivate the appropriate components (that implement se-

curity capabilities). This approach provides security

capabilities as-a-Service to fulﬁl the SLOs included

in the signed Security SLA.

Each identiﬁed security capability is implemented

by an appropriate security mechanism able to cover a

set of pre-deﬁned security controls. In SPECS, a se-

curity mechanism is a piece of software dedicated to

implementing security features on the target service.

The information associated with a security mecha-

nism is included in the mechanism’s metadata, pre-

pared by the mechanism’s developer and includes all

information needed to automate the security mecha-

nism’s deployment, conﬁguration and monitoring.

Cloud-automation tools, such as Chef

, can be

used to automatically implemented the security capa-

bilities required in the SLA through the Enforcement

services.

https://www.chef.io/chef/

Providing Security SLA in Next Generation Data Centers with SPECS: The EMC Case Study

141

4.1.3 Monitoring

In the Enforcement phase, the appropriate monitor-

ing components are also conﬁgured and activated.

The activation of monitoring components includes the

launching of services and agents that are able to mon-

itor the speciﬁc parameters included in the Security

SLA. These services and agents, which may be repre-

sented by existing monitoring tools integrated within

the framework, generate data that is collected and

processed by the SPECS Monitoring module (Casola

et al., 2015).

Under speciﬁc conditions, the Monitoring module

generates monitoring events, which are further pro-

cessed to verify whether they reveal a violation of the

SLA or indicate a possible incoming violation. As a

consequence, if a violation occurs, corrective counter-

measures may be adopted consisting of reconﬁguring

the service being delivered, taking the appropriate re-

mediation actions, or notifying the the End-User and

renegotiating/terminating the SLA.

5 THE SPECS ENHANCED ngDC

STORAGE TESTBED

A core objective of the SPECS framework is to de-

ploy cloud services with user deﬁned security capa-

bilities guaranteed by a Security SLA. SPECS allows

the End-user to express the desired capabilities in the

Security SLA using a reference standard for guid-

ance and clarity (e.g. Cloud Control Matrix (CCM)

3.0(CSA, 2015) and NIST SP 800-53 (NIST, 2013)).

This section describes the testbed used to integrate

SPECS and ViPR, the ViPR usage model and the ad-

ditional security features delivered as-a-Service and

guaranteed by a Security SLA.

5.1 Physical and Software Testbed

The core components of the architecture, illustrated in

Figure 6, are: (i) ESXi server; (ii) Cisco Switch; (iii)

VMAX array; (iv) Management Server.

The management server, running Windows Server

2008 R2 (W2K08), is used to set up and manage the

network to which the VMAX array is connected. Ad-

ditionally, EMCs SMI-S provider

must be installed

on this system in order to enable the management of

VMAX via ViPR. The VMAX array itself consists of

the VMAX controller and two VMAX bays equipped

with 800 disks for a total available storage space of

https://community.emc.com/docs/DOC-19629

200TB. ESXi

is VMware’s bare-metal hypervi-

sor that virtualizes servers and is installed directly on

top of the physical server, partitioning resources into

multiple virtual machines. These core components

are connected via high speed ﬁber channel managed

by a Cisco switch.

Figure 6: Physical Architecture.

In addition to these core components, the follow-

ing technologies are also used to support the test en-

vironment:

• VCenter Server: Provides a centralized and exten-

sible platform for managing virtual infrastructure.

• EMC ViPR Controller: Storage automation soft-

ware

• Chef: Powerful automation platform that is able

to automate the conﬁguration, deployment, and

management of VMs across different networks.

Figure 7 illustrates the high level conﬁguration of

the Chef Server alongside the SPECS core provider.

In this conﬁguration, Chef is used to install and con-

ﬁgure applications, and to deploy VMs. Services se-

lected from the list of available Chef cookbooks via

SPECS are deployed from the Chef Server as a Chef

recipe, launching a VM preconﬁgured to execute the

security service.

Figure 7: Software Architecture.

As can also be seen in Figure 7, the ViPR con-

troller is run as a virtual appliance (vApp) on the

https://www.vmware.com/products/

vsphere-hypervisor

CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science

142

ESXi server with the VMAX storage in the back-end

supplying the physical resources. The ViPR vApp is

deployed using the 3+2 conﬁguration ﬁle for redun-

dancy purposes (available from EMC support

). In

this deployment conﬁguration, ﬁve virtual machines

are used to run the vApp, with two VMs able to fail

without affecting the availability of the vApp.

5.2 ViPR Storage Service

The ViPR controller, accessible as a web application

running as a vApp on an ESXi server, offers End-

users virtual pools of storage resources using the as-

a-Service model. Using the ViPR REST API to ex-

ecute commands, the SPECS framework can access

all the functionality available through the controller.

Furthermore, this enables SPECS to allocate storage

services with additional security controls that can be

negotiated, for example Business Continuity Manage-

ment and Operational Resilience (BCR-01, BCR-09

and BCR-11 in the CCM control framework).

While it is possible for the End-user to set up secu-

rity features for the storage directly via ViPR, this re-

quires signiﬁcant technical and security expertise and

should be restricted to IT or Security administrators.

In contrast to this, the SPECS negotiation interface is

intuitive to personnel without speciﬁc administration

expertise, enabling them to select security require-

ments. Furthermore, the End-user can monitor the

enforcement of the security metrics over which the

SLOs have been deﬁned.

For the evaluation and testing of the SPECS

framework, the security control baselines deﬁned in

NIST SP 800-53 were selected to provide a standard-

ised mapping to the security mechanisms offered by

SPECS. This paper presents different security con-

trols across categories such as Access Control, Iden-

tiﬁcation and Authentication, Physical and Environ-

mental Protection, and System and Information In-

tegrity.

Table 1 shows some of the possible mappings used

between the NIST security controls and the security

features for Cloud storage services, expressed through

security metrics and their description.

On selecting the EMC ngDC SPECS application

from the SPECS portal, the End-user can choose from

different service conﬁguration parameters categorised

into the following Security Capabilities:

• Secure Storage Capabilities: Security capabili-

ties added to support services offered by storage

providers.

https://support.emc.com/downloads/32034 ViPR

Table 1: Sample mappings between NIST Security Controls

and SPECS Security Features.

Metric Description NIST

mapping

RAID

Level(s)

Deﬁnes the RAID level

the volumes in the vir-

tual pool will consist of.

SA-2, SC-

6, CP-9,

CP-10,

SI-17

SAN

Multi-

Path

The number of paths that

can be used between a

host and a storage vol-

ume.

SC-6, SI-

Data

Geoloca-

tion

Deﬁnes in which data

center the virtual stor-

age and its copies are lo-

cated.

PE-17,

PE-18,

PE-20,

SI-12

Max

Mirrors

Deﬁnes the Maximum

number of data storage

mirrors.

SC-5, SC-

6, SI-13

• Availability Capabilities: Capabilities providing

redundancy and business continuity in the event

of security incidents involving the storage service.

Each type of capability is responsible for a spe-

ciﬁc security aspect and is associated with a group of

security controls. For example, the Availability Ca-

pabilities are associated with SC-6 (Resource Avail-

ability) and SI-17 (Fail-Safe Procedures) as deﬁned

in NIST SP 800-53 for Security and Privacy Controls

(FORCE and INITIATIVE, 2013).

On selecting the type of capabilities required for

the storage service, the End-user is then presented

with the available capabilities under the selected cate-

gories. For example, the Availability Capabilities cat-

egory includes capabilities such as RAID level, High

Availability, Maximum Snapshots, etc.

Once the End-user has speciﬁed their require-

ments, the SPECS portal will display an overview of

the capabilities requested for the storage service, as

shown in Figure 8. This form displays the metric,

the associated value and the importance weight asso-

ciated with each.

Once the Agreement has been submitted and

signed, the Implementation function (of the Enforce-

ment module) implements the signed SLA by making

a series of requests via the ViPR REST API to set up

the storage service according to the requested capa-

bilities and security mechanisms. Furthermore, dur-

ing the implementation phase, the ViPR monitoring

agents are conﬁgured according to the metrics speci-

ﬁed in the SLA. The SPECS Monitoring module can

then make requests via ViPR’s REST API to contin-

uously check the status of the allocated storage and

verify if any SLA violation occurs.

Providing Security SLA in Next Generation Data Centers with SPECS: The EMC Case Study

143

Figure 8: SLA negotiated through SPECS.

Once the enforcement process is successfully

completed, the resources are available through the

ViPR administration interface. The capabilities re-

quested through SPECS are reﬂected in the provi-

sioned storage.

6 RELATED WORK

The drive towards the adoption of SLAs by CSPs is an

important initiative in strengthening the trust in ser-

vices. SLA management frameworks like SLA@SOI

(Theilmann et al., 2008) associate services with an

SLA, detect SLA violations and are even able to re-

cover from them. Many research projects, like Con-

trail (Morin, 2011), Optimis

and mOSAIC include

SLAs in their framework.

ENISA ((Dekker, 2012)(Catteddu,

2011)(Marnix Dekker, 2011)) outlined the need

for a Security SLA that offers clear guarantees with

respect to the security provided by CSPs to services.

Projects like CUMULUS (Pannetrat et al., 2013),

A4Cloud (Pearson, 2011), SPECS (Rak et al., 2013),

SLAReady

, SLALOM

, MUSA (Rios et al., 2015)

are actively working on this topic, attempting to

clearly model and represent security into an SLA.

Security and compliance issues in the ngDC are a

primary concern due to the reliance on traditional se-

curity models that have not adapted to virtualization.

In the ngDC, and cloud computing as a whole, new,

signiﬁcant risks have been introduced to IT services.

Organisations who traditionally hosted workloads and

data in internal data centres running on their own in-

frastructure, now face a loss of visibility and control.

The trust boundaries that were clearly established in

physical infrastructures are now blurred as virtualized

http://www.optimis-project.eu/

http://www.sla-ready.eu/

http://slalom-project.eu/

resources are increasingly used.

These new security issues have spawned several

research activities into novel solutions to address the

security of data in the cloud. For example, Nithi-

avathy proposes a framework to check the data in-

tegrity on CSP using homomorphic token and dis-

tributed erasure-coded data (Nithiavathy, 2013). Al-

ternatively, a secure multi-owner data sharing scheme

for cloud users using group signature and broadcast

encryption was proposed in (Marimuthu et al., 2014).

Other works focus on the security of the com-

munication between distributed Data Centers such as

(Talpur et al., 2015), that proposed a security frame-

work to manage a large number of secure connections

implemented using Kinetic

and Pyretic

as a cen-

tralized middleware tool. Each of these solutions ad-

dress a part of the problem, but do not offer a platform

on which the CSC can combine security requirements

that can be addressed in different application layers,

or give formal assurance to End-user through SLAs.

7 CONCLUSIONS

Next generation Data Centers provide a signiﬁcant

evolution how resources, such as storage, network and

compute, can be dynamically provisioned. The ngDC

offers the possibility to virtualize resources and dy-

namically pool them according to customer needs in

an Infrastructure-as-a-Service provisioning model.

To achieve broader adoption, datacenter cus-

tomers need more guarantees about the security lev-

els provided, creating the need for tools to negotiate

security requirements and to be able to monitor their

enforcement. This paper has investigated the poten-

tial of integrating the SPECS platform with the com-

mercially available ViPR storage solution from EMC.

This integration was shown to offer security as-a-

service, enhancing the security provided by the ViPR,

and guaranteed by a Security SLA.

In particular, this paper has illustrated, with a real

case study, how it is possible to negotiate, enforce and

monitor a Security SLA in a ngDC architecture. Po-

tential future directions from this work should focus

on the deﬁnition of new security metrics that can be

easily measured and monitored in order to provide

new security capabilities to a storage infrastructure

and enable a CSP to enrich its security service offer-

ings.

http://resonance.noise.gatech.edu/

http://frenetic-lang.org/pyretic/

CLOSER 2016 - 6th International Conference on Cloud Computing and Services Science

144

ACKNOWLEDGEMENTS

This research is partially supported by the EC FP7

project SPECS (Grant Agreement no. 610795).

REFERENCES

Casola, V., Benedictis, A. D., and Rak, M. (2015). Secu-

rity monitoring in the cloud: An SLA-based approach.

In 10th International Conference on Availability, Re-

liability and Security, ARES 2015, Toulouse, France,

August 24-27, 2015, pages 749–755.

Casola, V., Benedictis, A. D., Rak, M., and Villano,

U. (2014). Preliminary Design of a Platform-as-a-

Service to Provide Security in Cloud. In CLOSER

2014 - Proceedings of the 4th International Con-

ference on Cloud Computing and Services Science,

Barcelona, Spain, April 3-5, 2014., pages 752–757.

Catteddu, D. (2011). Security & resilience in governmental

clouds. Technical report, CSA.

CSA (2015). Cloud controls matrix v3.0.

CSCC (2012). The cscc practical guide to cloud service

level agreements. Technical report, CSCC.

Davidson, E. A. (2013). The Software-Deﬁned-Data-Center

(SDDC): Concept Or Reality? [VMware].

De Benedictis, A., Rak, M., Turtur, M., and Villano, U.

(2015). Rest-based sla management for cloud ap-

plications. In Enabling Technologies: Infrastructure

for Collaborative Enterprises (WETICE), 2015 IEEE

24th International Conference on, pages 93–98.

Dekker, M. (2012). Critical cloud computing a ciip perspec-

tive on cloud computing services. Technical report,

ENISA.

EC (2011). Unleashing the potential of cloud computing in

europe. Technical report, EC.

FORCE, J. T. and INITIATIVE, T. (2013). Security and

privacy controls for federal information systems and

organizations. NIST Special Publication, 800:53.

International Organization for Standardization (2014).

ISO/IEC NP 19086-1. Information Technology–

Cloud computing–Service level agreement (SLA)

framework and technology–Part 1: Overview and con-

cepts.

Marimuthu, K., Gopal, D. G., Kanth, K. S., Setty, S.,

and Tainwala, K. (2014). Scalable and secure data

sharing for dynamic groups in cloud. In Advanced

Communication Control and Computing Technologies

(ICACCCT), 2014 International Conference on, pages

1697–1701. IEEE.

Marnix Dekker, G. H. (2011). Survey and analysis of secu-

rity parameters in cloud slas across the european pub-

lic sector.

Morin, C. (2011). Open computing infrastructures for elas-

tic services: contrail approach. In Proceedings of

the 5th international workshop on Virtualization tech-

nologies in distributed computing, pages 1–2. ACM.

NIST (2013). SP 800-53 Rev 4: Recommended Security

and Privacy Controls for Federal Information Systems

and Organizations. Technical report, NIST.

Nithiavathy, R. (2013). Data integrity and data dynamics

with secure storage service in cloud. In Pattern Recog-

nition, Informatics and Mobile Engineering (PRIME),

2013 International Conference on, pages 125–130.

IEEE.

Pannetrat, A., Hogben, G., Katopodis, S., Spanoudakis, G.,

and Cazorla, C. (2013). D2.1: Security-aware sla

speciﬁcation language and cloud security dependency

model. technical report, certiﬁcation infrastructure for

multi-layer cloud services (cumulus).

Pearson, S. (2011). Toward accountability in the cloud. In-

ternet Computing, IEEE, 15(4):64–69.

Rak, M., Suri, N., Luna, J., Petcu, D., Casola, V., and Vil-

lano, U. (2013). Security as a service using an sla-

based approach via specs. In IEEE, editor, Proceed-

ings of IEEE CloudCom Conference 2013.

Rios, E., Iturbe, E., Orue-Echevarria, L., Rak, M., and Ca-

sola, V. (2015). Towards self-protective multi-cloud

applications - MUSA - a holistic framework to sup-

port the security-intelligent lifecycle management of

multi-cloud applications. In CLOSER 2015 - Pro-

ceedings of the 5th International Conference on Cloud

Computing and Services Science, Lisbon, Portugal,

20-22 May, 2015., pages 551–558.

Talpur, S. R., Abdalla, S., and Kechadi, T. (2015). To-

wards middleware security framework for next gen-

eration data centers connectivity. In Science and In-

formation Conference (SAI), 2015, pages 1277–1283.

IEEE.

Theilmann, W., Yahyapour, R., and Butler, J. (2008). Multi-

level sla management for service-oriented infrastruc-

tures. In Proceedings of the 1st European Conference

on Towards a Service-Based Internet, ServiceWave

’08, pages 324–335, Berlin, Heidelberg. Springer-

Verlag.

Providing Security SLA in Next Generation Data Centers with SPECS: The EMC Case Study

145