Alessandro Carrega

Cutting the energy bills of Internet Service Providers and telecoms through power management

The energy consumption of the Information and Communication Technology (ICT) sector has been increasing recently; this sector is estimated to account for 2% of the total energy consumption. An even more aggressively increasing trend is the volume of Internet traffic and the number of connected devices. Thus, reducing the energy needs of the Internet is recognised as one of the main challenges that the ICT sector will have to face in the near future to reduce its overall energy footprint. Introducing energy-efficient techniques, both at the device level and the network level, is required.The main goal of this work is to quantitatively evaluate the potential energy savings from applying energy-efficient techniques, while examining the trade-off between network performance and the achieved energy savings.We introduce a categorisation of the energy-aware design space, focusing on the existing techniques in the device data plane, and contribute an analytical framework to represent the impact of energy-aware technologies and solutions for network devices. Our energy profile model represents the diverse energy-aware states of the network devices and is applied over two reference scenarios, one of a large-scale Telco (Telecom Italia) and one of a medium size Internet Service Provider (GRNET), to evaluate the impact of each energy-aware technology and the energy savings potential at the Home, Access, Metro/Transport and Core parts of each network.The results show the estimates of energy savings exceed 60% in many cases, while maintaining the same quality of service as in the energy-agnostic case.

Green Networking With Packet Processing Engines: Modeling and Optimization

With the aim of controlling power consumption in metro/transport and core networks, we consider energy-aware devices able to reduce their energy requirements by adapting their performance. In particular, we focus on state-of-the-art packet processing engines, which generally represent the most energy-consuming components of network devices, and which are often composed of a number of parallel pipelines to “divide and conquer” the incoming traffic load. Our goal is to control both the power configuration of pipelines and the way to distribute traffic flows among them. We propose an analytical model to accurately represent the impact of green network technologies (i.e., low power idle and adaptive rate) on network- and energy-aware performance indexes. The model has been validated with experimental results, performed by using energy-aware software routers loaded by real-world traffic traces. The achieved results demonstrate how the proposed model can effectively represent energy- and network-aware performance indexes. On this basis, we propose a constrained optimization policy, which seeks the best tradeoff between power consumption and packet latency times. The procedure aims at dynamically adapting the energy-aware device configuration to minimize energy consumption while coping with incoming traffic volumes and meeting network performance constraints. In order to deeply understand the impact of such policy, a number of tests have been performed by using experimental data from software router architectures and real-world traffic traces.

Green network technologies and the art of trading-off

In this contribution, we focus on energy-aware devices able to reduce their energy requirements by adapting their performance. We propose an analytical model to accurately represent the impact of green network technologies (i.e., low power idle and adaptive rate) on network- and energy-aware performance indexes. The model has been validated with experimental results, performed by using energy-aware software routers and real-world traffic traces. The achieved results demonstrate how the proposed model can effectively represent energy- and network-aware performance indexes. Moreover, also an optimization procedure based on the model has been proposed and experimentally evaluated. The procedure aims at dynamically adapting the energy-aware device configuration to minimize energy consumption, while coping with incoming traffic volumes and meeting network performance constraints.

A Closed-Form Model for the IEEE 802.3az Network and Power Performance

We propose an analytical model able to accurately estimate both power consumption and network performance indexes of Energy Efficient Ethernet (EEE) links working at the three available speeds under various traffic load patterns and packet size distributions. The model is sufficiently flexible and accurate to consider different traffic parameters; among others, the packet size distribution, the average burst inter-arrival rate, the burst size distribution, etc. With relatively low complexity, since it addresses stationary queue behavior, the analysis allows obtaining the average energy consumption of the link and, unlike previous works, the mean latency time experienced by incoming packets in closed form, without any upper or lower bound approximations. This aspect makes the model suitable to be adopted in optimization frameworks for network design and control purposes. The numerical results of the model are validated against measurements on a real test bench.

Theoretical and technological limitations of power scaling in network devices

The largest part of routers and switches, today deployed in production networks, has very limited energy saving capabilities, and substantially requires the same amount of energy both when working at full speed or when being idle. In order to dynamically adapt such energy requirements to the real device work load, current approaches foster the introduction of low power idle and power scaling primitives in entire devices, internal components and network interfaces. Starting from these considerations, we focus on power scaling, and we propose an analysis of the theoretical and technological limitations in adopting such kind of mechanisms. Thus, our contribution is twofold. On one hand, we performed several tests to identify the technological limitations in a software router based on off-the-shelf hardware, which already includes such capabilities. The results achieved show that the power scaling allows a linear trade-off between consumption and network performance, but the time to switch between two power states may cause a non negligible service interruption. On the other hand, regarding the theoretical limitations, we consider the trade-off between the benefit in dynamically adapting the power states within short time-scales and the overhead needed to choose and select the new power state.

Applying traffic merging to datacenter networks

The problem of reducing energy usage in datacenter networks is an important one. However, we would like to achieve this goal without compromising throughput and loss characteristics of these networks. Studies have shown that data-center networks typically see loads of between 5%-25% but the energy draw of these networks is equal to operating them at maximum load. To this end we examine the problem of reducing the energy consumption of datacenter networks by merging traffic. The key idea is that low traffic from N links is merged together to create K ≤ N streams of high traffic. These streams are fed to K switch interfaces which run at maximum rate while the remaining interfaces are switched to the lowest possible rate. We show that this merging can be accomplished with minimal latency and energy costs (less than 0.1W total) while simultaneously allowing us a deterministic way of switching link rates between maximum and minimum. We examine the idea of traffic merging using three different datacenter networks - flattened butterfly, mesh and hypercube networks. In addition to analysis, we simulate these networks and utilizing previously developed traffic models we show that 49% energy savings are obtained for 5% per-link load while we get 20% savings for a 50% load for the flattened butterfly and somewhat lower savings are obtained for the other two networks. The packet losses are statistically insignificant and the maximum latency increase is less than 3μs. The results show that energy-proportional datacenter networks are indeed possible.

An analytical model for designing and controlling new-generation green devices

In this paper, we focus on energy-aware devices able to reduce their energy requirements by adapting their performance. We consider the device to be able to save energy through two main energy-aware primitives, namely, low power idle and power scaling. In such an environment, we propose a novel and original model for accurately representing how the joint usage of the previously cited primitives can impact on both energy consumption and network performance. As shown by the results achieved, the proposed model can be effectively applied in order to design and control energy-aware hardware of next-generation network devices.

Exploiting novel software development paradigms to increase the sustainability of data centers

The application of effective energy management strategies in data centers is often hindered by the substantial conflict between the interests of cloud users and infrastructure owners. As a matter of fact, cloud users require that the service level they are paying for is tightly met, whereas data center owners try to cut down their operational expenses. In this paper, we propose a novel consolidation algorithm that exploits emerging software development paradigms. Our approach enables cloud users to indicate their willingness to apply energy saving mechanisms to some of their virtual resources, hence giving infrastructure managers the ability to apply more efficient workload consolidation and to switch their hardware to very low-power states. The result is an optimal trade-off between energy consumption and performance.

A Middleware for Mobile Edge Computing

Telecom operators have recently started to deploy massive computing and storage resources at the very edge of their access networks, hence evolving their infrastructures into large, distributed, and capillary computing environments, capable of placing applications very close to users and terminals and well-suited to effectively fulfill the challenging requirements for delay-sensitive applications. Edge computing is the emerging paradigm that pervasively brings computing in the environment, by shifting processing from data centers to the network edge, allowing a large class of applications in growing fields, like Big Data and the Internet of Things, to be deployed in a very effective way. In this paper, we propose middleware for running applications over heterogeneous environments, made of telecom networks and legacy data centers. We briefly review an emerging architecture for edge computing and describe an integrated solution for developing and deploying modular applications in an automatic way.

A network-centric architecture for building the cloud continuum

The growing interest in distributed, context-aware and data-sensitive applications is pushing the evolution of computing infrastructures from centralized to distributed models, which could effectively tackle the execution of complex software frameworks over geographical scale. The concept of cloud continuum that extends computing infrastructures beyond the data center boundary will require new architectural paradigms that overcome the evident limitations intrinsic in mere cloud federation and that enable effective and efficient interaction between the cloud and the physical environment. In this paper, we discuss why and how telecommunication networks could be the most effective infrastructure to create a distributed, pervasive, carrier-grade cloud continuum, by acting as the core federation paradigm for the dynamic and flexible composition of data centers, networks and IoT platforms.