Virtualization of Heterogeneous Wireless-Optical ... - IEEE Xplore

TZANAKAKI LAYOUT_Layout 1 8/1/13 4:09 PM Page 155

TOPICS IN OPTICAL COMMUNICATIONS

Virtualization of Heterogeneous Wireless-Optical Network and IT Infrastructures in Support of Cloud and Mobile Cloud Services Anna Tzanakaki and Markos P. Anastasopoulos, Athens Information Technology Georgios S. Zervas, Bijan Rahimzadeh Rofoee, Reza Nejabati, and Dimitra Simeonidou, University of Bristol

ABSTRACT This article proposes a next generation ubiquitous converged infrastructure to support cloud and mobile cloud computing services. The proposed infrastructure facilitates interconnection of fixed and mobile end users with data centers through a heterogeneous network integrating optical metro networks, based on time shared optical network technology, and wireless access networks, based on Long Term Evolution access technology. To support the infrastructure as a service paradigm, the proposed architecture adopts the concept of virtualization across the technology domains involved. Planning of virtual infrastructures considering jointly the presence of all network technology domains and IT resources is proposed, with the aim to offer globally optimized virtual infrastructures (VIs) in terms of energy consumption and resource requirements. The holistic VI planning approach proposed ensures allocation of the required resources across all technology domains to support not only the volume of service requests, but also their specific characteristics such as end users’ mobility. Our modeling results clearly show that both the volume and characteristics of services have a direct impact on the energy consumption and resource requirements of all the technology domains of the planned VIs.

INTRODUCTION As the availability of high-speed Internet access is increasing at a rapid pace, distributed computing systems that are able to support a large variety of existing and upcoming applications are gaining increased popularity. Over the last decade, large-scale computer networks supporting both communication and computation were extensively employed to run distributed applications that deal with enterprise IT services, internet control processes, web content presentation, media services, file sharing etc. Traditionally, data center (DC) constellations were locally

IEEE Communications Magazine • August 2013

installed at the customer’s site. However, this solution suffers from severe limitations, including: • Heavy instrumentation on IT infrastructure and personnel needs • Difficulties in adapting the IT infrastructure requirements to the rapidly increasing or unpredictable computing needs • Increased operational and maintenance costs for the IT infrastructures • Limited reliability as a failure of the central computers may affect the operation of the entire system In response to these observations, the current trend is on-demand delivery of infrastructures, applications, and business processes in a commonly used, secure, scalable, and computer based environment over the Internet for a fee [1], known as cloud computing. In cloud computing, subscription-based access to infrastructures may be provided to users that is referred to as infrastructure as a service (IaaS). Cloud computing architectures comprise a variety of hardware and software components communicating with each other through a high-performance network infrastructure. On the other hand, cloud computing services need to be supported by specific IT resources that may be remote and geographically distributed between themselves and the end users, requiring connectivity through a very highcapacity network with increased flexibility and dynamicity. A strong candidate to support these needs is optical networking due to its carriergrade attributes, its abundant capacity, its energy efficiency, and recent technology advancements including dynamic control planes and so on. In this environment, cloud computing services, hosted by DC infrastructure components, are emerging as an essential element of the enterprise IT infrastructure. Also, recently the concept of mobile computing is gaining increased attention, as it aims at supporting the additional requirement for ubiquitous access of mobile end users to computing resources. Mobile computing imposes the requirement that portable devices

0163-6804/13/$25.00 © 2013 IEEE

155


Data centers

TSON edge Metro optical nodes network 2

3 Data centers

TSON frame

Mobile users

1 Wireless access

LTE - wireless access

Fixed users

IP packets from the LTE BS are forwarded to the TSON edge nodes, 1 The encapsulated in Ethernet packet. TSON edge node processes the Ethernet traffic, and after converting it 2 The to TSON frames, forwards it to the destined TSON edge node. extracts the Ethernet packets from the TSON frames, and directs 3 TSON them to the data center.

Figure 1. Physical infrastructure.

run standalone applications and/or access remote applications via wireless networks, moving computing power and data storage away from the mobile devices to remote computing resources, according to the mobile cloud computing (MCC) paradigm [2]. Advantages offered by MCC include: • Dynamic on-demand provisioning and utilization of resources • Improving data storage capacity and processing power by accessing remote computation resources through wireless access • Improving reliability by resilience mechanisms available in the cloud • Extending battery lifetime through computation offloading techniques enabling migration of large computations from resource-limited devices (i.e., mobile devices) to resourceful machines (i.e., servers in clouds) It is predicted that cloud and MCC services are emerging as one of the fastest growing business opportunities for Internet service providers and telecom operators. To enable this emerging business opportunity, there is a need for converged infrastructures supporting heterogeneous network domains integrating wireless access and high-capacity optical networks to interconnect IT resources with end users in support of the IaaS concept as well as the cloud and the MCC paradigm. Existing best effort Internet solutions cannot inherently and effectively address this requirement as they do not support convergence and interoperability of dynamic heterogeneous broadband and mobile network technologies, seamless transparent end-to-end service provisioning mechanisms with quality of service (QoS) guarantees, and optimized schemes to support energy and resource efficiency. On the other hand, infrastructure virtualization is key in cloud computing, as it allows efficient sharing of physical resources (i.e., computing, storage, and network), while providing deterministic QoS through infrastructure isolation. Virtualization techniques have been commercially exploited by many cloud service providers, and it is expected that future cloud services will be based on globally distributed DCs interconnected through network infrastruc-

156

tures deploying cross-domain coordinated virtualization. This article proposes a next generation ubiquitous converged infrastructure to support the cloud and MCC services of the future. This infrastructure facilitates the interconnection of DCs with fixed and mobile end users through a heterogeneous network integrating optical metro and wireless access network technologies. The proposed architecture will address the diverse bandwidth requirements of future cloud services by integrating advanced optical network technologies offering fine (sub-wavelength) switching granularity with state-of-the-art wireless Long Term Evolution (LTE) access network technology supporting end-user mobility through wireless backhauling. To support the IaaS paradigm as well as the diverse and deterministic QoS needs of future cloud and mobile cloud services, the proposed architecture adopts the concept of virtualization across the technology domains. The proposed virtual infrastructure (VI) planning is taking a holistic approach that considers jointly the presence of all network technology domains and the IT resources to offer globally optimal solutions in terms of specific objectives such as energy consumption and resource requirements. This holistic approach also ensures allocation of the required resources across all technology domains to support not only the volume of service requests, but also their specific characteristics such as mobility of end users. Our modeling results identify trends and trade-offs relating to resource requirements and energy consumption levels of the infrastructure across the various technology domains that are directly associated, with the volume and nature of the services supported.

ARCHITECTURE The infrastructure model proposed is based on the IaaS paradigm and aims at providing a technology platform interconnecting geographically distributed computational resources that can support a variety of cloud and mobile cloud services. To support the IaaS paradigm, the proposed architecture comprises a heterogeneous network integrating optical metro and wireless access networks interconnecting DCs, and adopts the concept of physical resource virtualization across the technology domains. Virtualization facilitates the sharing of physical resources among various virtual operators and services, introducing new business models and enabling new exploitation opportunities for the underlying physical infrastructures.

PHYSICAL INFRASTRUCTURE In the proposed solution end users will access and share computational resources remotely on an on-demand basis in accordance with the cloud computing paradigm. The overall physical infrastructure is illustrated in Fig. 1. Wireless Access and Backhauling Solutions — In this study, the wireless domain of the physical infrastructure is described through a heterogeneous topology comprising a cellular LTE system [3] for the wireless access part and a collection of wireless microwave links for the inter-



connection of the LTE-enabled based stations and edge nodes of the optical metro network solution. LTE is among the prime wireless access cellular technologies in the fourth generation (4G) standard. LTE is anticipated to offer a theoretical net bit rate capacity of up to 100 Mb/s per sector in the downlink and 50 Mb/s per sector in the uplink if a 20 MHz channel is used. These data rates can be further increased if multipleinput multiple-output (MIMO) technology is adopted. At the same time LTE marks the transition from a circuit-switched to an all-IP network architecture enhanced with greatly improved QoS characteristics such as low packet transmission delays (smaller than 5 ms for small packets), fast and seamless handover from one cell to another supporting up to 350 km/h for high-speed vehicular communications scenarios, operation with different bandwidth allocations (scalable bandwidth up to and including 40 MHz, while operation in wider bandwidths, e.g., up to 100 MHz, is also possible). Furthermore, LTE can support a wide range of services and performance metrics (e.g., real-time and nonreal-time streaming, conversational, and interactive services with low or high delay as well as background) in a wide range of environments such as indoor, urban, and rural. Optical Metro Network Solution — To support the metro network segment, we propose the use of a wavelength-division multiplexed (WDM) optical metro network technology referred to as the time shared optical network (TSON) [4]. TSON is designed and implemented as a novel frame-based, time multiplexing metro network solution, offering dynamic connectivity with fine granularity of bandwidth. TSON is a contentionless solution through the deployment of a central resource allocation engine of route, wavelength, and time assignment, responsible for setting up the sub-wavelength paths. TSON has been developed in the EU project MAINS [4], and inherently facilitates a distributed cloud environment where the IT resources are located at the metro region in order to reduce the backbone traffic and meet the required quality of experience (QoE). It is predicted that in the next 10 years a single internal metro flow will occupy a fraction of a wavelength, and the deployment of subwavelength networks can enhance the metro network utilization. On the other hand, sub-wavelength bandwidth granularity supporting short-lived connections will facilitate fast time-to-service delivery (a few milliseconds), low end-to-end delay, and multiple levels of guaranteed QoS. In this context, TSON seems to be well suited to support the cloud and mobile cloud services of the future. TSON’s functionality is illustrated in Fig. 1. TSON was originally designed to support nomadic cloud services where the users could remotely access virtual PCs on distributed DCs. When the users move, the virtual PCs may also move to other DCs as appropriate. In the proposed scenario where TSON is integrated with the wireless access LTE network supporting mobile users, mobility is accommodated by TSON through reallocation of services to differ-


ent DCs depending on the relevant location of the end users. From a technology perspective this is possible as the TSON metro nodes allow flexible network services through a number of attributes, including: • Fine granularity of 100 Mb/s per 10 Gb/s wavelength channels • Flexible resource allocation and enhanced bandwidth utilization • Flexible and fine resource virtualization capability • Modular node architecture for scalability and integration with other systems (e.g., elastic and inelastic optical networks) • QoS guarantees required for distributed cloud and mobile cloud services with respect to delay and contention

A key challenge is that of cross-domain and cross-technology virtualization for the creation of infrastructure slices including IT as well as optical and wireless network resources, in support of the IaaS paradigm.

VIRTUALIZATION A key challenge on which this article focuses is that of cross-domain and cross-technology virtualization for the creation of infrastructure slices including IT as well as optical and wireless network resources, in support of the IaaS paradigm. This will involve the abstraction of physical resources into logical resources that can then be assigned as independent entities to different VIs, and shared by a variety of virtual operators and end users. The objective is to implement dynamically reconfigurable unified VIs over the underlying converged optical and wireless network segments interconnecting IT resources. These VIs are aimed at satisfying the VI operators’ requirements and end users’ needs, while maintaining cost effectiveness and other specific requirements such as energy efficiency. This process involves the identification of the optimal VI that can support the required services in terms of both topology and resources, and includes mapping of the virtual resources to the physical resources. Figure 2 illustrates a virtual topology formed over a physical topology and the associated mapping of resources between the two layers, physical and virtual. In this article we propose to apply virtualization and perform VI planning, taking a holistic approach that considers jointly the heterogeneous network technology domains and the IT resources comprising the physical infrastructure to offer a globally optimal solution in terms of specific objectives such as energy consumption and resource requirements. Through this joint consideration of all technology domains, it is ensured that sufficient resources are allocated across all infrastructure segments to support not only the volume of service requests but also their specific characteristics such as mobility of end users and required QoS guarantees.

OPTIMIZATION ACROSS MULTITECHNOLOGY DOMAINS As described already, we propose to plan virtual DC infrastructures (VDCs) considering jointly the presence of all network technology domains and the IT resources incorporated in the physical infrastructure with the aim of offering an optimized VDC in terms of specific objectives. More specifically, the objective of the proposed

157


The physical infrastructure under consideration is

7

1

S1

6 S4

S5

2

S2

4

5

described through a topology comprising a cellular LTE system

S3

Virtual topology 2 4 7

6 8

microwave links for the interconnection

E1

3

collection of wireless

12

5 13

1

for the wireless access domain and a

S6

3

heterogeneous

13

11

10 E2

9

E3

Physical topology

of the LTE-enabled base stations

Data center

Wireless backhaul link

S1-S6 Sources

Virtual links

Unidirectional optical link

E1-E3 Edge nodes

Figure 2. Virtualization over heterogeneous network infrastructures. optimization scheme is to identify the virtual topology and determine the virtual resources required to implement a VDC over the available wireless access, wireless backhaul, optical metro network, and DC physical resources optimizing a specific set of requirements such as energy consumption (operational cost) and resource requirements. This VDC not only will meet customers’ specific needs, which may be either known in advance or predicted based on history observations, but will also satisfy the VI provider’s requirements for minimum operational and capital expenditures. In a highly dynamic heterogeneous environment, the problem of optimal VI planning is complex since information regarding the position and application requirements of mobile devices, the performance of the optical and wireless network domains, as well as the availability of resources in the network and the DCs are uncertain. In order to assess the performance and requirements of this type of VIs we have developed for the first time, an optimization scheme suitable for VDC planning taking into account both the time variability and uncertainty of services over an integrated IT and heterogeneous network infrastructure suitable for cloud and mobile cloud services. The physical infrastructure (PI) under consideration is described through a heterogeneous topology comprising a cellular LTE system for the wireless access domain and a collection of wireless microwave links for the interconnection of the LTE-enabled base stations. In terms of wired technologies, computing resources are interconnected through the TSON WDM metro network with frame-based sub-wavelength switching granularity incorporating active nodes that will also interface the wireless micro wave links supporting backhauling from the wireless access segment. This infrastructure is suitable to support traffic that is generated by both traditional cloud and mobile cloud applications. In this context, traf-

158

fic demands corresponding to traditional cloud applications are generated at randomly selected nodes in the wired domain and need to be served by a set of IT servers. Mobile traffic, on the other hand, is generated at the wireless access domain and in some cases needs to traverse a hybrid multihop wireless access/backhaul solution before it reaches the IT resources through the optical metro network. The performance of cloud and mobile cloud computing services is determined not only by the IT resources used to service them, but also by the details of the associated network resources. In the formulation of our model, it is assumed that the granularity of network demands is a portion of wavelength (e.g., l/100), while the IT locations at which the services will be handled are not specified and are of no importance to the services themselves. Therefore, identification of the suitable IT resources that will support the services is part of the optimization output. In the general case, the VI planning problem should be solved taking into account a set of constraints that guarantee the efficient and stable operation of the resulting infrastructures. The following constraints are considered: 1) Every demand has to be processed at a single IT server. This allocation policy reduces the complexity of implementation. 2) The planned VI must have sufficient capacity for all demands to be transferred to their destinations. 3) The capacity of each link in the VI should be realized by specific PI resources. A critical issue to be taken into account is the distinction between traffic that arises from fixed and mobile devices. For example, referring to Fig. 2, virtual link 7 can be realized through specific resources in the physical path consisting of the optical links 12 and 13. Similarly, for a scenario with fixed wireless users, virtual link 1 can be realized using a combination of the physical paths 1-2-5 and 3-4-5 in the wireless domain. On the other hand, for a scenario with mobile users,



Minimize F = ∑ kg (u g ) + ∑ wl (ul ) + ∑ ∑ σ sr (hd ) g

l

s r

The costs considered in our modeling are related to the energy consumption of the infrastructure, which can be directly associated with the operational expenditure (OPEX) of the planned VI. For the data centers, the power consumption model presented in [6] has been adopted where for a service rate of x Mb/s, the corresponding power consumption is defined as ps + asx, where the subscript indicates the server, p s is the baseline consumption in the idle state, and as is the slope of the load-dependent consumption. In the optical network domain, the cost for each optical link comprises the energy consumed by each lightpath due to transmission and reception of the optical signal, optical amplification at each fiber span, and switching [7]. The switching power consumption of the TSON solution is based on actual laboratory measurements. Furthermore, for the wireless backhaul, the power consumption model presented in [8] has been adopted where wireless backhaul links are treated as a collection of wireless microwave links of 100 Mb/s capacity and a power dissipation of 50 W each. Thus, for a given average backhaul requirement per base station, cbh Mb/s, the backhaul power consumption is 0.5 W/Mb/s. Finally, in the wireless access domain, the power consumption model of the LTE-enabled base station presented in [9] has been adopted.


22 20 Average number of wavelengths

virtual link 1 should be realized using a combination of the physical paths 1-2-5 and 3-4-5, as well as a portion of the resources in 6-7-10 and 8-9-10. The accurate estimation of the resources that should be reserved by the VI provider (VIP) to ensure that seamless end-to-end service provisioning is based on the mobility model adopted, the size of the LTE cells, and the traffic model that is adopted. In the ideal scenario, a seamless handoff for a mobile device can be 100 percent guaranteed if the VIP reserves an equivalent amount of resources to all its neighboring cells. However, a more sophisticated approach would be to relate the reserved resources in the neighboring LTE cells with the handoff probability. In the current article, the amount of resources that are leased in the wireless domain by the VIP is assumed to be an increasing function of the handoff probability [5]. To address this issue, initially the handoff probability is calculated, and then a portion of the resources are used. 4) Finally, the planned VI must have adequate IT server resources such as CPU, memory, and disk storage to support all requested services. The objective of our formulation is to minimize the total cost during the planned time frame of the resulting network configuration, which consists of the following components: • kg: the cost for operating capacity ug of the PI link g in the optical domain • wl: the cost for operating capacity ul of the PI link l in the wireless domain • ssr: the total cost of the capacity resource r of IT server s for processing the volume of demand hd

18

WDM (network resource optimization) TSON (network resource optimization) WDM (IT resource optimization) TSON (IT resource optimization)

16 14 12 10 8 6 4 2 0

1000

6000 11,000 16,000 Demands/source (Mb/s)

21,000

Figure 3. Time-shared optical network architecture.

NUMERICAL RESULTS To investigate the performance of the proposed VI design scheme across the multiple domains involved, the architecture illustrated in Fig. 2 is considered: the lower layer depicts the PI and the layer above depicts the VIs. For the PI, a macro-cellular network with regular hexagonal cell layout has been considered, similar to that presented in [9], consisting of 12 sites, each with 3 sectors and 10 MHz bandwidth, operating at 2.1 GHz. The inter-site distance (ISD) has been set to 500 m to capture to scenario of a dense urban network deployment. Furthermore, 2 ¥ 2 MIMO transmission with adaptive rank adaption has been considered, while the users are uniformly distributed over the serviced area. As indicated in [9], each site can process up to 115 Mb/s, and its power consumption ranges from 885 to 1087 W, under idle and full load, respectively. For the computing resources, three Basic Sun Oracle Database Machine Systems have been considered where each server can process up to 28.8 Gb/s of uncompressed flash data, and its power consumption ranges from 600 to 1200 W under idle and full load, respectively [10]. For the metro network, the TSON [4] solution has been adopted assuming a single fiber per link, 4 wavelengths per fiber, wavelength channels of 10 Gb/s each, minimum bandwidth and granularity 100 Mb/s, and maximum capacity of the network 40 Gb/s. Finally, cloud and MCC traffic is generated at randomly selected nodes in the optical metro network and LTE base stations, and needs to be served at the three IT servers. Note that in order to study how the end-user mobility and traffic parameters affect the utilization and power consumption of the planned VI, the service-to-mobility factor [5] is introduced. The service-to-mobility factor is defined as the fraction of the service holding time over the cell residence time. It is assumed that the cell residence time is exponentially distributed with parameter, and the service holding time is Erlang distributed with parameters (m, n).

159


The efficiency of the proposed TSON solution compared to a traditional WDM metro network supporting wavelength switching granularity is investigated using the physical topology illustrated in Fig. 2 under two scenarios: • Optimizing in terms of network resource requirements • Optimizing in terms of IT resource requirements Figure 3 clearly indicates the resource efficiency achieved by TSON with regard to the required wavelengths in the optical metro network segment. Figure 4 illustrates the total power consumption of the converged infrastructure (wireless access, wireless backhaul, optical network, and IT resources) when applying the proposed approach optimizing for network or IT resources. Comparing these two schemes, it is observed that the energy-aware VI design consumes significantly lower energy (lower operational cost) to serve the same amount of demands than the closest IT

8000

Power consumption (W)

7000

DC Optical network Wireless backhaul Wireless access

6000 5000 4000 3000 2000 1000 0

Network

IT

Network

7650 Mb/s

IT

Network

10,710 Mb/s

IT

Network

13,770 Mb/s

IT

153,00 Mb/s

Total load (mobile cloud/cloud traffic = 0.02)

Total power consumption (W)

Figure 4. Impact of traffic load on power consumption when the VI is planned for energy or distance.

9000 8000 7000 6000 5000 4000 0

1

2

3

Service-to-mobility factor

4

5

6000 5000 4000 3000 2000 Average 1000 demands/source (Mb/s)

Figure 5. Impact of mobility and traffic load on the total power consumption (m = 5).

160

scheme, providing an overall saving of the order of 37 percent for low traffic demand. This is due to the fact that in the former approach fewer IT servers are activated to serve the same amount of demand. Given that the power consumption required for the operation of the IT servers is dominant in this type of infrastructure, switching off the unused IT resources achieves significant reduction of energy consumption. Furthermore, it is observed that for both schemes the average power consumption increases almost linearly with the number of demands. However, the relative benefit of the energy-aware design decreases slightly with the number of demands when approaching full system load. It is also observed that the wireless access technology is responsible for 43 percent of the overall power consumption, while the optical network consumes less than 7 percent of the total power. To further investigate how the service characteristics affect the optimum planned VI, the impact of both traffic load and end-user mobility on the total power consumption (operational cost) of the VI is studied. Figure 5 shows that the total power consumption increases for higher end-user mobility, as expected. More specifically, when mobility is higher (lower service-to-mobility factor), additional resources are required to support the VI in the wireless access domain. However, it is interesting to observe that this additional resource requirement also propagates in the optical metro network and the IT domain. The additional resource requirements, across the various infrastructure domains, are imposed in order to ensure availability of resources in all domains involved (wireless access and backhauling, optical metro network, and DCs) to support the requested services and enable effectively seamless and transparent end-to-end connectivity between mobile users and the computing resources. Finally, Fig. 6 illustrates the impact of mobility on the additional resources required across the multitechnology domains for various optimization schemes. Again, it is observed that in order to achieve seamless end-to-end connectivity between end users and computing resources, additional resources should be reserved to support the VI. For fast moving mobile terminals and services with high duration (low service-tomobility factor), more than 150 percent additional optical network resources (compared to the static scenario) should be allocated to the VI to ensure seamless service provisioning across the various network segments. For high mobility, the cases where services are realized through different TSON gateways are increased, and therefore, the resources in the optical domain should be increased.

CONCLUSIONS A next generation ubiquitous converged infrastructure to support cloud and mobile cloud computing services was proposed. This infrastructure facilitates the interconnection of fixed and mobile end users with DCs through a heterogeneous optical metro and wireless access network. A physical infrastructure integrating state-of-the-art wireless LTE access


network technology, supporting end-user mobility and an advanced optical metro solution based on TSON technology, offering the required flexibility and fine bandwidth granularity to match the characteristics of the cloud and mobile cloud services is assumed. To support the IaaS paradigm, the proposed architecture adopts the concept of resource virtualization across the technology domains involved. Planning of VIs, taking a holistic approach that considers jointly all network technology domains and the IT resources incorporated in the physical infrastructure, was proposed. This approach offers globally optimized VIs and ensures allocation of the required resources across all technology domains, addressing both the volume of service requests and their specific characteristics such as mobility of end users. Our modeling results show that there is a direct association between the characteristics of the services to be supported and the requirements in terms of resources and power consumption of the planned VI that spans across the various technology domains. More specifically, for higher end-user mobility, additional resources are required not only in the wireless access but also in the optical metro network and IT domains. These additional resources ensure seamless and transparent end-to-end connectivity between mobile users and computing resources.

ACKNOWLEDGMENT This work has been partially supported by the EC through the STREP project CONTENT (INFSO-ICT-318514) and the FP7 PIANO+ ADDONAS project.

REFERENCES [1] M. A. Rappa, “The Utility Business Model and the Future of Computing Systems,” IBM Sys. J., vol. 43, no. 1, 2004, pp. 32–42. [2] H. Dinh et al., “A Survey of Mobile Cloud Computing: Architecture, Applications, and Approaches,” Wireless Commun. Mobile Comput., Oct. 2011 [3] 3GPP TS 36.213, “Technical Specification Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA); Physical Layer Procedures.” [4] G. S. Zervas et al., “Time Shared Optical Network (TSON): A Novel Metro Architecture for Flexible MultiGranular Services,” Opt. Express 19, B509-B514, 2011. [5] Y. Fang and I. Chlamtac, “Analytical Generalized Results for Handoff Probability in Wireless Networks,” IEEE Trans. Commun., vol. 50, no. 3, Mar. 2002, pp. 369–99. [6] V. Valancius et al., “Greening the Internet with Nano Data Centers,” Proc. 5th Int’l Conf. Emerging Networking Experiments and Technologies, pp. 37–48. [7] A. Tzanakaki et al., “Energy Efficiency in Integrated IT and Optical Network Infrastructures: The GEYSERS Approach,” Proc. IEEE INFOCOM ’11, Wksp. Green Commun. and Net., pp. 343–48. [8] A. J. Fehske, P. Marsch, and G. P. Fettweis, “Bit per Joule Efficiency of Cooperating Base Stations in Cellular Networks,” Proc. IEEE GLOBECOM, Dec. 2010. [9] G. Auer and V. Giannini, “Cellular Energy Efficiency Evaluation Framework,” Proc. IEEE VTC-Spring, Hungary, May 2011. [10] “Sun Oracle Data Base Machine Data Sheet,” http://www.oracle.com/us/solutions/datawarehousing/039569.pdf.

BIOGRAPHIES ANNA TZANAKAKI [SM] ([email protected]) is an associate professor at Athens Information Technology (AIT) and is leading the Network Design and Services research group. She is coauthor of over 150 international publications, co-inventor of


Additional resources required (%)


160% 140% 120% 100% 80% 60% 40% 20% Optical network Wireless access Wireless backhaul DC

0% Network IT Network IT Network IT 3 Network 2 1 0 Service-to-mobility factor

IT

Figure 6. Impact of mobility on the additional resources required across the multitechnology domains for various optimization schemes (mobile cloud traffic = 50 Mb/s, cloud traffic 500 Mb/s/source).

1 granted and 11 published patents, and a technical referee for various journals and conferences. Dr. Tzanakaki serves on several conference Technical Program Committees. Her research interests include network architectures, technologies, and protocols in support of the Future Internet. M ARKOS A NASTASOPOULOS received his Diploma degree in electrical and computer engineering (5-year program), M.S. in financial engineering, and Dr. Eng. degree from the National Technical University of Athens. Currently, he is a research engineer in the NDS group of AIT. His research interests lie in the areas of optical and wireless communications networks, mobile and distributed computing, network design, and management. He is a member of the Technical Chamber of Greece. BIJAN RAHIMZADEH ROFOEE received his M.Sc. degree with distinction with the IEEE UK/RI Communication Prize from the University of Essex, United Kingdom, in 2009. Since then he was a research assistant at the University of Essex and afterward at the University of Bristol. He has been involved in a number of national and EU funded projects focusing on next generation optical networks, sub-wavelength and super-wavelength communication systems, high performance opto-electornic modules, and reconfigurable and programmable networks. G EORGIOS Z ERVAS is a lecturer in optical and high-performance networks at the University of Bristol, United Kingdom. He received M.Eng. and Ph.D. degrees, and was a research fellow and lecturer at the University of Essex. He has participated in several EC and national projects. His research interests include flexible, cognitive, and software/hardware defined optical networks. He is a coauthor of over 100 publications and a TPC member of international conferences. He is involved in IETF and OGF, and has filed two patents. R EZA N EJABATI is a lecturer at the University of Bristol. He obtained his Ph.D. from the University of Essex, where he has been a lecturer and an RCUK fellow. His area of research is in disruptive new Internet technologies with focus on application of high-speed network technologies, design, and control issues for software defined, service-oriented, and programmable networks, cross-layer network design and architecture, as well as network architecture and technologies for e-science and cloud computing. DIMITRA SIMEONIDOU is a professor of high-performance networks at the University of Bristol. She is a leading academic focusing on optical networks, future Internet research, and experimentation and grid/cloud computing, and a pioneer in transport software defined networking. She is chairing several international conferences/committees, and is a co-author of over 350 publications, 11 patents, and several major contributions to standards. She is Chair of the EU FP7 projects cluster on Converged and Optical Networks (CaON).

161