=Paper=
{{Paper
|id=Vol-1203/EES-paper2
|storemode=property
|title=A Look at Energy Efficient System Opportunities with Community Network Clouds
|pdfUrl=https://ceur-ws.org/Vol-1203/EES-paper2.pdf
|volume=Vol-1203
|dblpUrl=https://dblp.org/rec/conf/ict4s/FreitagSKNBGV14
}}
==A Look at Energy Efficient System Opportunities with Community Network Clouds==
https://ceur-ws.org/Vol-1203/EES-paper2.pdf
A Look at Energy Efficient System Opportunities
with Community Network Clouds
Felix Freitag∗ , Leila Sharifi∗‡ , Amin M. Khan∗‡ , Leandro Navarro∗ , Roger Baig† , Pau Escrich† , Luı́s Veiga‡
∗ Department of Computer Architecture. Universitat Politècnica de Catalunya. Barcelona, Spain
{felix, lsharifi, mkhan, leandro}@ac.upc.edu
† Fundació Privada per la Xarxa Lliure, Oberta i Neural Guifi.net. Mas l’Esperana, 08503 Gurb, Catalonia
{roger.baig, pau.escrich}@guifi.net
‡ Instituto Superior Técnico. Universidade de Lisboa. INESC-ID Lisboa. Lisbon, Portugal
{luis.veiga}@inesc-id.pt
Abstract—Community networking is an emerging model of
a shared communication infrastructure in which communities
of citizens build and own open networks. Community networks
offer successfully IP-based networking to the user. In addition,
some hosts are connected to the network nodes in order to
provide network management and end user services. Recently,
clouds have been proposed for community networks. Some
research projects such as Clommunity have started deploying
computational infrastructure to enable cloud computing within
community networks. In this paper we propose different options
for such community clouds to contribute to energy efficient
systems, in particular regarding cloud-based services and in
relation to Smart Grid. Further discussion and interaction with
the research initiatives on energy efficient systems should identify
the most promising approach and outline possible ways for
implementation.
Index Terms—community networks; cloud computing
Fig. 1. Guifi.net nodes and links in the area around Barcelona
I. INTRODUCTION features which may contribute to energy efficient cloud system
Community networks are user-driven communication net- design and implementation.
works which often originated from the lack of Internet con- • Distributed and heterogeneous hardware: Cloud hardware
nectivity in rural areas. There are several community-owned in community clouds is contributed by the users, hardware
networks in the range of 500 to 20,000 nodes in Europe such can be heterogeneous and it is at the users’ premises.
as FunkFeuer, AWMN, Guifi.net, Freifunk and many more • Decentralized service management: The network and
worldwide. Most of them are based on Wi-Fi technology (ad- cloud services are managed by the users. Services may
hoc networks or IEEE 802.11a/b/g/n access points in the first be shared among users to build collaborative applications.
hop, long-distance point-to-point Wi-Fi links for the trunk • Community-driven services: The potential of the com-
network), but also a growing number of optic fibre links have munity cloud is to deploy services of local interest.
started to become deployed [1]. Guifi.net in Spain (Figure 1) Awareness of energy-efficiency may be brought to the
is probably the largest community network worldwide and it is users, which could drive a fast roll-out of services to
where the cloud deployment takes place which we are currently support energy-efficiency.
undertaking. We describe in the following section first the characteristics
The community cloud we consider here are cloud deploy- of the cloud deployment that we have started in the Guifi.net
ment in community networks: A cloud hosted on community- community network. Secondly, we sketch some proposals to
owned computing and communication resources providing benefit from community network cloud infrastructure to build
services of local interest. The concept of community clouds energy efficient systems.
has been introduced in its generic form before, e.g., [2], [3],
as a cloud deployment model in which a cloud infrastructure II. CLOUDS IN COMMUNITY NETWORKS
is built and provisioned for an exclusive use by a specific
community of consumers with shared concerns and interests, In this section we explain some components of the cloud we
owned and managed by the community or by a third party or have deployed in Guifi.net community network, which presents
a combination of both. a real case for our proposals for extending towards energy-
Such community cloud built in community networks have efficiency in section III.
Copyright c papers’ authors. Copying permitted only for private and academic purposes. This volume is published and copyrighted by its editors.
Fig. 3. Cloud resource at a community network node
Fig. 2. Conceptual schema of microclouds in a community wireless network
cloud deployment at time of this writing can be seen in the
Clommunity project’s Wiki2 .
A. Topology
C. Cloud Management Platform
Cloud in community networks needs to fit into the condi-
tions of how these networks are built, governed and grown. Since the different local groups providing such microclouds
Regarding their growth, one important aspect is the topology are independently organized, the cloud management platforms
of community networks, which determines the scope of the (CMP) that are used are also heterogeneous. In the commu-
options how the cloud infrastructure can be integrated. In nity clouds we have deployed we use mainly Proxmox and
Figure 2 some node types of a community wireless network are OpenStack as CMP. The reason for using Proxmox is that
depicted. The picture shows some typical community nodes within the community network, there is already some positive
with a router and some server or clients attached to it. Some usage experience, and the installation and operation of Proxmox
clients nodes are shown that are connected to the access point is relatively easy compared to other CMPs. OpenStack on
(AP) of a super node. In addition, however, these community the other hand, is popular as a powerful customizable cloud
nodes have cloud resources attached to them which are part of platform, supported by a large user community, though not
the community cloud. In addition, depending on the topology of within community networks.
the wireless network and the social structure of the community, D. Software Distribution
several local community clouds may appear which manage
We provide a community cloud GNU/Linux distribution,
locally the cloud resources belonging to a certain zone defined
codenamed Cloudy, aimed and designed for building clouds in
geographically or by its social networks. We call such local
community networks. This distribution contains the platform
clouds microclouds. A set of microclouds are interconnected
and application services of the community cloud model. Cloudy
though the wireless links of the super node backbone network.
is deployed on the cloud resources that form the microcloud
and over the different microclouds in the community network,
B. Hardware as illustrated in Figure 4.
In Figure 3 an example of the indoor hardware of a cloud I I I . S U P P O RT O F C O M M U N I T Y C L O U D S F O R
node is shown. In this case a small Jetway device is used ENERGY EFFICIENT SYSTEMS
as cloud resource. A UPS keeps the node running in case of
In this section we sketch a set of proposals on how to benefit
power failures. It is connected over Ethernet to the outdoor
from community clouds for building energy efficient systems.
community network node. This cloud nodes represents the
case of low-end cloud resources such as home gateways, that A. Energy-efficient resource allocation
users provide to the cloud. These devices have been deployed Community clouds are envisioned as a set of federated
through Community-Lab1 [4] and can be integrated in the microclouds. Microclouds are composed of heterogeneous
Guifi network management platform [5]. Other cloud nodes devices, which can range from resource-constrained home
we have deployed are several Dell OptiPlex 7010 desktops. In gateways up to server class desktops, even small crowd-funded
addition, some Alix boards and Intel Galileo boards have been data centres can be imagined. Depending on the service-level
deployed for specific purposes. While the number of deployed agreements of the requested cloud service, different cloud
cloud resources in Guifi is in constant evolution, the status of resources may be able to fulfil the requirements.
1 http://community-lab.net/ 2 http://wiki.clommunity-project.eu/testbed:start
reduce the need of infrastructure that utility providers need to
deploy in the last mile of the Smart Grid [6]. For further
exploration of this scenario, the requirements of Smart Grid
platforms need to be assessed. The community cloud which we
consider does principally allow to deploy customized PaaS or
integrate additional services into the Cloudy distribution, which
may allow easier exploration than through closed commercial
environments.
I V. C O N C L U S I O N
The appearance of clouds in community networks can be
witnessed nowadays. Such clouds consist of user contributed
infrastructures which enable cloud-based services; the cloud
Fig. 4. Cloudy distribution deployed in microclouds hardware is often located at the user premises. Such community
clouds offer unique features compared to commercial data
centre type clouds, such as distributed and heterogeneous
If an energy model of the different classes of cloud hardware
hardware, user-managed services and services of community
available in the community clouds is available, the most energy
interest. This paper sketches some opportunities which com-
efficient cloud infrastructure may be selected to which this job
munity clouds could offer to contribute to energy efficient
should be allocated. An initial proof of concept of such a
systems. These opportunities are described towards energy-
system could be studied on a single heterogeneous microcloud.
efficient resource allocation, clouds made out of low-power
The level of federation support of the used CMP may allow
devices, and supporting the deployment of the Smart Grid.
extending such energy-efficient resource allocation over several
microclouds. AC K N OW L E D G M E N T
In this scenario, the heterogeneity of community cloud This work was supported by the European Community
hardware plays an essential role. Such energy-efficient resource Framework Programme 7 FIRE Initiative projects CLOMMU-
allocation may be less applicable in commercial cloud systems NITY, FP7-317879, and Community Networks Testbed for the
where the hardware is more homogeneous. Future Internet (CONFINE), FP7-288535.
B. Cloud computing with low-power devices REFERENCES
Different research initiatives, e.g. the Montblanc project3 , [1] “Guifi.net new sections of fiber deployed to the farm, 2012.” [Online].
consider building energy-efficient cloud computing infrastruc- Available: http://en.wikinoticia.com/Technology/internet/122595
tures with low-power devices. [2] P. Mell and T. Grance, “The NIST Definition of Cloud Computing,”
NIST Special Publication, vol. 800, no. 145, 2011. [Online]. Available:
Community cloud contributors are expected to target cloud http://csrc.nist.gov/publications/nistpubs/800-145/SP800-145.pdf
hardware deployed on their premises which has a low energy [3] A. Marinos and G. Briscoe, “Community Cloud Computing,” in Cloud
Computing, First International Conference, CloudCom 2009, ser. Lecture
consumption, for being able to run in a 24/7 mode without Notes in Computer Science, M. Jaatun, G. Zhao, and C. Rong, Eds.
noticeable electricity consumption increase. Therefore, such a Beijing, China: Springer Berlin Heidelberg, Dec. 2009, vol. 5931, pp.
cloud scenario built upon many distributed low-power devices 472–484.
[4] B. Braem, R. Baig Viñas, A. L. Kaplan, A. Neumann, I. Vilata i Balaguer,
could be materialized by a user-provisioned cloud infrastruc- B. Tatum, M. Matson, C. Blondia, C. Barz, H. Rogge, F. Freitag, L. Navarro,
tures in community clouds. J. Bonicioli, S. Papathanasiou, and P. Escrich, “A case for research with
In order for such a scenario to be explored, CMPs need and on community networks,” ACM SIGCOMM Computer Communication
Review, vol. 43, no. 3, pp. 68–73, Jul. 2013.
to be evaluated towards their capacity to support hypervisor [5] J. Jiménez, R. Baig, P. Escrich, A. M. Khan, F. Freitag, L. Navarro,
solutions which may not require hardware supported virtualiza- E. Pietrosemoli, M. Zennaro, A. H. Payberah, and V. Vlassov, “Supporting
tion. Several low-power devices have already been deployed cloud deployment in the Guifi.net community network,” in 5th Global
Information Infrastructure and Networking Symposium (GIIS 2013).
in Community-Lab [4], making the hardware to explore this Trento, Italy: IEEE, Oct. 2013.
scenario mainly available. [6] L. Sharifi, F. Freitag, and L. Veiga, “Combining Smart Grid with Commu-
nity Clouds: Next Generation Integrated Service Platform,” in 5th IEEE
C. Smart Grid support International Conference on Smart Grid Communications (SmartGrid-
Comm 2014). Venice, Italy: IEEE, Nov. 2014.
Smart Grid scenarios build upon advanced metering and
communication infrastructures, which enable to take more
energy-efficient decisions on power consumption. For such
scenarios, computational infrastructure is needed to process
large amounts of data generated by electrical meters and com-
bine it with additional contextual information. Such processing
could leverage on community cloud hardware and services, and
3 http://www.montblanc-project.eu/