Novel Architecture for Multimedia Distribution based on Content Aware Networking
Eugen Borcoci
Telecommunication Dept.
University Politehnica of Bucharest
Bucharest, Romania
e-mail: eugen.borcoci@elcom.pub.ro
Daniel Negru
CNRS-LaBRI Lab.
University of Bordeaux, France
e-mail:daniel.negru@labri.fr
Christian Timmerer
Department of Information Technology
Faculty of Technical Sciences
Klagenfurt University, Austria
e-mail: christian.timmerer@itec.uni-klu.ac.at
Abstract—This paper proposes a novel virtual Content Aware
Network (CAN) layer as a part of a full layered architecture,
focused, but not limited, to multimedia distribution with
quality of services assurance. The overall system is based on a
flexible cooperation between providers, operators and endusers, enabling users to access the offered multimedia services
in various contexts and also to become private content
providers. The paper introduces the main concepts and
architecture for the main virtual network layer (CAN),
exposing its role and interfaces among overall system layers.
This work is a part of the starting effort inside of a new
European research FP7 IP project, ALICANTE.
Keywords: content aware nertworking, network aware
applications, quality of services, multimedia distribution, Future
Internet
I.
INTRODUCTION
It is widely accepted today that Future Internet (FI)
should solve a lot of issues, not satisfactorily fulfilled by the
current one. This affects all architectural layers, being related
to new service paradigms, flexibility, scalability, security and
trust, user-friendly behavior, flexible business model and
roles of entities, coping with heterogeneity of lower layer
network
technologies,
generalized
mobility,
etc.
Revolutionary, evolutionary or hybrid approaches are
proposed, [1], [2], [4]. Among them the concept of network
virtualization is considered as a main way for Internet
evolution to allow diversification of the next architectures,
[1] , [2], [5] in a flexible manner.
On the other side from the end user perspective the FI is
defined by the set of services it provides, [3]. The end user
sees FI as a tool providing easy access to various type of
information in real and/or non-real time (general data, voice,
video, music, multimedia), search engines, different services
(e.g. access, personal communication, business and
entertainment services), or social networking platforms.
The “content orientation” is a main FI paradigm. Digital
multimedia services and networked media content are
playing an increasing role in citizens’ economic, social and
cultural life, touching all domains as industry, administration,
business, education, culture and entertainment. This trend is
recognized also by the European commission, which defined
the “Objective ICT-2009.1.5: Networked Media and 3D
Internet” in the FP7 Call 4, 2009, [4], [6]. Here, new targets
are defined as “Content aware networks (CAN) and network
aware applications (NAA)”. We can see here a new approach
in which the neutrality of the network is no more supposed,
but intelligence and a higher degree of coupling to the upper
layers are embedded in the network nodes. Based on
virtualization, the network can offer enhanced transport and
adaptation capable services. It is hoped that in such a way,
architectures and technologies for converged and scalable
networking and delivery of multimedia content and services
can be achieved. Dynamic optimization is desired, with
policies taking into account the content and adaptation needs,
the user contexts, requirements and social relational network.
This should be done for a variety of contents and services
including home management and various applications,
locations and mobility scenarios. The FI should enable
multiple user roles as content producer, user or manager.
The work of this paper is a part of the starting effort
inside of a new European research FP7 IP project, “MediA
Ecosystem Deployment Through Ubiquitous Content-Aware
Network Environments”, ALICANTE, [19]. The project
proposes a new “Media Ecosystem”, gathering a mass of
existing, but also new potential content creators and media
Service Providers (SP), essentially stemming from customers
(active users). Such a “Media Ecosystem”, by analogy with
the ecology or business counterparts, can be characterized by
inter-working environments to which the actors belong and
through which they collaborate, in the networked media
domain. These environments are: User Environment (UE), to
which the End-Users belong; Service Environment (SE), to
which the Service and Content Providers belong; Network
Environment (NE), to which the Network Providers belong.
By “Environment”, we understand here a generic and
comprehensive name to emphasize a grouping of functions
defined around the same functional goal and possibly
spanning, vertically, one or more several architectural (sub-)
layers. This name is used to characterize its broader scope,
w.r.t the term “layer”. By “Service”, if not specified
differently, we understand here high level services, as seen at
application/service layer.
The above environments are nowadays present in real
deployments, but actually there is not sufficient collaboration
between them. The User context is not taken into
consideration by the Service (or Content) Provider delivering
the service (content), which, as a consequence, could not be
capable of adapting the service (content) to the capabilities of
the user. The current architectures do not include exchange
of content-based and network-based information between the
Network layers and upper layers, implying that CAN and
Network-Aware services/applications have difficulties to
emerge. This neutral network service considered many years
as a good principle, proves nowadays to be a weak solution,
especially if one considers the new multimedia
communications and their increasing importance in the FI.
The paper is organized as follows: the Section II presents
samples of related work. The Section III defines the main
concepts and overall ALICANTE architecture. The Section
IV is the main part of the paper; it is focused on the CAN
layer functionality and interfaces. Conclusion, open issues
and future work is shortly outlined in the Section V.
II.
RELATED WORK
The content aware networking (CAN) and network aware
applications (NAA) approach is a new mode to design the
layered architecture, breaking the classic TCP/IP and OSI
stack concepts. Traditionally, the way in which contents are
generated, processed, and distributed are seen as separated
from the transport itself. The challenge and question is
whether in the context of new requirements of multimedia
and combined service flows, one can enable better
interactions without losing modularity of the architecture. In
such a context, both content-aware networking (adjusting
network resource allocation based on limited examination of
the nature of the content) and network-aware content
processing (adjusting the way contents are processed and
distributed based on limited understanding of the network
condition) are of high interest both for research communities
and industry, in the overall framework of re-thinking the
architecture of the Future Internet.
In [7] it is considered that CAN and NAA can offer a way
for evolution of networks beyond IP. One can benefit from
routing based on content and/or context and embed business
rules into high speed low latency networks. The capability of
a content adaptive network awareness to offer joint
optimization of video transmission is analyzed in [8]. The
CAN/NAA approach, can naturally lead to an user-centric FI
and telecommunication services, as it described in [9]. The
work [10] discusses the content adaptation issues in the FI as
a component of CAN/NAA approach. Naturally, the
CAN/NAA approach supports better QoE/QoS capabilities of
the future networks, [11], [15]. The architecture can be still
more rich if to content awareness we add add contextawareness, [12], [16]. New content-aware cooperative
multiple access protocol for packetized voice is proposed in
[13].
On the other side the high amount of packet header
processing necessary to be embedded in the CAN network
elements, raises serious problems similar to Deep Packet
Inspection techniques, [14]. So, it is needed to find new
methods to minimize the processing tasks in the high speed
routers.
Another contribution that CAN/NAA approach can bring
to solve the current networking problems is related to the
P2P traffic overload of the global internet, [17], [18]. The
application layer traffic optimization (ALTO) problem
studied at IETF can be helped by the cooperation between
the CAN layer and the upper layer, following in principle the
ideas mentioned in [17], [18], but solved via CAN approach.
However there is currently no complete and open
architecture for multimedia distribution, able to
accommodate both – all, current and future needs of
multimedia content oriented services on one side, and
flexible, scalable and efficient usage of network transport
resources over heterogeneous networking technologies on the
other side. Therefore an open field for research in this
domain exists.
III.
ALICANTE SYSTEM ARCHITECTURE
The ALICANTE architecture promotes advanced
concepts: Content-awareness to the NE; User Contextawareness to the SE, and adapted services/content to the
End-User for his/her best service experience while being a
consumer and/or producer.
A. Novel Virtual Layers
Two novel virtual layers are proposed on top of the
traditional Network layer, i.e. CAN layer for network level
packet processing and a Home-Box (HB) layer for the actual
content delivery. Innovative components, instantiating the
CAN are called Media-Aware Network Elements (MANE).
They are actually CAN-enabled “routers” and associated
managers, offering together content-aware and context-aware
Quality of Service/Experience, security, and monitoring
features, in cooperation with the other elements of the
ecosystem.
The upper layer, i.e., the SE, uses information delivered
by the CAN layer and enforces network-aware applications
procedures, in addition to user context-aware ones. The novel
proposed Home-Box (HB) entity is a physical and logical
entity located at End-User's premises and gathering ContextAware, Content-Aware and Network-Aware information
essential for realizing the big picture. Associated to the
architecture, there exists an open, metadata-driven,
interoperable middleware for the adaptation of advanced,
distributed media resources to the user's preferences and
heterogeneous contexts enabling an increased Quality of
Experience. The adaptation will be deployed at both HB and
CAN layers making use of scalable media resources. Finally,
the validation of the project architecture and results will be
done in a large-scale international pilot, in preparation for
bringing it to the market.
B. Business Model
The architecture defined can support a set of actors,
representing a new business model, enough flexible and
capable to satisfy the Service/Content/Network Providers’
and End-Users objectives and needs, together with flexibility
in offering new possibilities in terms of their role as
consumers, providers and managers. The main business
actors/entities envisaged are the following:
Content Consumer (CC) or End-User (U) is an entity
(human plus a terminal or a process), which establishes a
contract with an SP for service/content delivery. These users
are the final target recipients of services. A CC might also be
an organization acting on behalf of individual End-Users.
Content Provider (CP) gathers/creates, maintains, and
distributes digital information. The CP owns/operates
network hosts (content sources) but it might not own any
networking infrastructure to deliver the content. The content
is offered to the CCs or SPs through Service Level
Agreements. There may be business relationships between
CPs and NPs to host or co-locate the content servers that
belong to CPs. The CC can also be a private CP.
Service Provider (SP) delivers to CCs high level services
and aggregates content obtained from multiple CPs. SPs may
not necessarily own a transport infrastructure, but rely on the
connectivity services offered by Network Providers (NPs), or
CAN Providers (CANP). CCs have interactions with SPs and
SPs are ultimately responsible for the service offered to
them. SPs may interact with each other in order to expand
their service base. SPs use the services of CPs and NPs, or
CANPs, through appropriate Service Level Agreements –
SLAs;
Network Provider (NP) traditionally offers connectivity
providing reachability between network domains/hosts. NPs
own and administer IP connectivity infrastructures. They
interact with each other for the purpose of expanding the
geographical span of the offered connectivity services.
Home-Box (HB) is a new ALICANTE business entity,
which can be partially managed by the SP, the NP and the
End-User. The HBs can cooperate with SPs in order to
distribute multimedia services (e.g. IPTV) in different modes
(e.g. P2P);
CAN Provider (CANP) is a new ALICANTE business
entity, seen as a virtual layer functionality provider. It is
actually an enhanced virtual NP (note that the additional
CAN functions are performed by network nodes and
therefore they are executed by the Network Provider). It
offers content-aware network services to the upper layer
entities.
C. Hierarchy of Functions
More specifically, ALICANTE proposes several new
features inside the Network/Service/User Environments in
order to enable the future generic “Media Ecosystem” that:
At network level, it realizes and offers to upper layers, a
rich and virtualized networked multimedia space, through an
all-IP prototype for environment, customizable for delivering
networked media content. The architecture is capable of:
- applying CAN concepts, to perform network/transport
intelligent content-aware processing (routing, dynamic
adaptation, security, etc.) for existing and future emerging
applications in a scalable, open and optimized way. This is
the main role of the new CAN layer;
-realizing distributed management and control in order to
customize the CANs as to respond to the upper layer needs,
including 1-1,1-n and n-m communications, and also allow
efficient network resource exploitation at network provider
level;
- performing cross layer optimizations between the
virtual CAN layer and upper layers, including but not limited
to peer-to-peer (P2P) approach. This optimization will be
possible due to network awareness capabilities of the upper
layers;
-extending CAN functionalities for achieving an efficient
collaboration with elements in the Service Environment,
enabling content-awareness and network-awareness;
At service/content level, it delivers enriched networked
media services and content, which can be efficiently
exploited by End-Users. The architecture new features are:
-elaborating a new approach for the delivery of services
which includes the HB as a new element in the service
distribution chain, capable of advanced functionalities
(service management and adaptation, user mobility,
security);
- creating a new virtual HB layer, composed of virtually
interconnected HB (in traditional distributed client/server
mode or P2P mode), capable of advanced provisioning of
service/content;
- dissociation of the Service/Content Providers’ roles and
capabilities and the Home-Box layer role and capabilities in
terms of service/content exploitation and delivery, which will
lead to the vision of their efficient cooperation;
- enhanced services: delivery - through the servers or
HBs, in various modes; discovery - the introduction of a new
type of component called Service Registry (SR); efficient
management; service composition - realization of a Service
Composition Engine supporting streaming media
applications;
-achieving collaboration with the User Environment and
with the CAN Network Environment.
At user level, it allows the users to consume and/or
generate content and exploit services delivered by
components of the SE. The new related architecture features
are:
- adding new dimensions to the user by giving him the
possibility to have several roles, such as: Content and Service
Consumer; Content and Service Provider; Content and
Service Manager;
- elaborating a User Profile to characterize the static and
dynamic parameters of the user and his context, in order to
be exploited by SE elements (HB) for the delivery of adapted
services;
- permitting any user to access/deliver/manage any
service/content on any device from anywhere and at any
time, thanks to a specific User/Service interface and QoE
monitoring tool at the user’s terminal side;
- achieving efficient collaboration with the SE, enabling
User Context-Awareness, for the best End-User experience.
At all levels, monitoring is performed in several points of
the service distribution chain and regulates a twofold
adaptation action, at the virtual HB Layer and at the virtual
CAN Layer.
IV.
CONTENT AWARE NETWORK LAYERED
ARCHITECTURE
The Figure 1 presents the overall architecture showing
the environments described in the previous sections and
emphasizing the CAN layer and physical perspective of the
system. The main interaction between the SE environment
and CAN layer are summarized by arrows denoted contentaware and respectively network-aware. The bottom part of
the figure shows a possible network infrastructure composed
of access part and several autonomous systems (AS).
Figure 1 The ALICANTE Architecture: details on Virtual CAN Layer
A. Vertical and Horizontal Layering
Two novel virtual layers are proposed on top of the
traditional Network layer, virtualizing the network nodes:
one layer for packet processing (CAN layer) and the other for
content delivery (Home-Box layer);
• Virtual Content-Aware Network (VCAN) layer offers an
enhanced support for packet payload inspection,
processing and caching in network equipment. It is
developed over traditional IP network/transport layer.
It will improve data delivery via classifying and
controlling messages in terms of content, application
and individual subscribers, improves QoS assurance via
Content-based routing and increases network security
level via content-based monitoring and filtering. In such
a way, content and application aware networks are
created to provide high levels of performance, EndUser experience, and to enable application and
subscriber-specific data forwarding. The specific
components in charge of creating this VCAN are the
Media-Aware Network Elements (MANE), i.e. the new
CAN routers, and the CAN managers.
• Virtual Home-Box layer is an upper layer, using CAN
services and
taking into account network-aware
information delivered upward by the CAN layer.
Thanks to this layer, inter-working with the User,
Service and Network Environments, one can elaborate
Network and
Context - Aware Applications and
deliver the necessary inputs to create Content-Aware
Networks. The adaptation, service mobility, security
and overall management of services and content are
being assured at this layer through a new specific
middleware proposed by the project, working in
conjunction with the other layers.
The interactions between the above mentioned two layers
establish together a powerful cross-layer optimization loop
providing End-Users with the best possible service
experience and optimizing the resource usage.
Seen from the traditional layered architecture perspective,
the system can be divided horizontally into three parallel
planes: Management, Control and Data Planes (MPl, CPl,
DPl) cooperating with each other. They are not represented
distinctly in the picture, which only emphasizes some
important interfaces and relationships, as presented below:
Management and Control Interfaces: The main
management and control entity in the VCAN layer is the
CAN Manager (CANMng). Corresponding to its roles, we
distinguish the following interfaces of CAN Manager with
the Virtual Home-Box layer: to advertise CANs and
negotiate their usage and to help the establishing of
connectivity relationships at Virtual HB layer based on e.g.
network related distance information. CANMng has also
interfaces to the lower network layer in oredr to negotiate
CANs and request their installation.
Data Plane Interfaces: The data plane interfaces
transport the packets between the VCAN layer and the
Virtual Home-Box layer in both directions. The downloaded
packets are especially marked by the application layer, so
that they can be associated with the correct CAN and to
allow their processing in the data plane of the CAN. To make
easier the processing task (performed by CAN nodes in the
data plane) and increase the overall performance, the
ALICANTE project may adopt the solution to pre-configure
the MANEs – CAN nodes – and allow them some freedom to
lightly decide on the packets themselves.
Each AS has one CAN Manager controlling that one or
several CANs are deployed in each domain. The CAN
Manager is the entity which plays the following roles: to
(re)define the CANs (according to the enhanced connectivity
service targeted) and perform all related actions to configure,
maintain and update CANs; to advertise and negotiate the
CAN usage with upper layers, using Service Level
Ageeements/Specifications
(SLA/SLS)
contracts;
to
communicate with other CAN managers in order to establish
multi-domain chains, again, using SLA/SLS contracts; to
communicate with its own intra-domain network resource
managers (Intra NRM). The IntraNRMs have the ultimate
authority upon the network provider resources, thus
conserving each domain’s independency.
B. The Content-aware Router
The content-aware router, (MANE), is an intelligent
network node. It takes into account the content type in order
to perform appropriate processing (filtering, routing,
adaptation, security operations, etc.) according to the content
properties (described by metadata or extracted by protocol
field analysis) and also depending on network properties and
its current status. The results of the content related
information analysis provide metrics, which allow deciding
the best strategy to adopt for the best content repurposing and
publishing methods. The MANE basic set of functions are:
Content-aware intelligent routing: the MANE will
decouple the routing process (higher level) by the
forwarding (lower level) and make intelligent routing, based
on
results
extracted
from
packet
fields’
analysis or content description metadata. Examples of
content-aware
functions
can
be
anycast and
routing based on publish-subscribe paradigms;
Content-aware QoS and resource allocation: the MANEs
will be able to implicitly deduct the QoS requirements of
different flows based on the flows content. The CAN layer
will monitor the current status of the CANs from the load
point of view. The MANE will maintain an aggregated image
of flows that they forward. For every recognized flow type,
an appropriate instance of CAN will be assigned depending
on the level of QoS guarantees and network status. Efficient
resource allocation and/or load balancing can be done in the
network depending on traffic types and QoS requirements,
by taking benefit from content awareness of MANE and
based on operator policies, in terms of resource allocation.
The CAN level will interact with the domain network
resource management in order to perform mapping onto
different L2/L3 QoS aware technologies (e.g. MPLS/Diffserv
or Carrier Ethernet). Dynamic re-allocation (not frequently,
in order to prevent instability) of the network resources
between different CANs can be done, to assure the flexibility
and efficiency of resource usage. The MANE has also a role
in the adaptation. The latter is actually deployed at different
points in the delivery chain: at the service creation, during
the transport by the CAN routers, and at the Home-Box site;
Specific Security issues: while keeping usage of current
and popular IPSec technologies at network level, the
ALICANTE project will exploit the possibility to include
content related information in dedicated fields. Thus, the
end-to-end communication remains encrypted and private
whilst the content-aware network can operate seamlessly.
Content awareness is a subject of in-depth packet
processing that spans across the traditional Layer 2
and Layer 3 approach, while the inspection can be driven, by
pre-defined policies, thus obtaining gain in speed and also
flexibility. Note the important aspect - related to privacy
issues - that packet inspection done by ALICANTE MANE
router is not at all applied to the content itself, but only on
metadata describing the content type, in order
to offer appropriate processing to these flows from several
points of view - quality of services, network resource
efficient usage, etc.
C. CAN Management Details
The CANMng has the following roles: defining the
CANs that may exist on its given IP core infrastructure,
based on domain policy information and requests on future
needs of customers; configuring, maintaining and updating
CANs; advertising and negotiating the CAN services offered
to upper layers; negotiating SLAs with other CAN Managers
in order to establish multi-domain CAN enabled chains;
cooperating with intra-domain manager to install, modify or
delete various virtual CANs in the network devices; helping
the establishment of relationships at HB layer, at requests of
HBs, e.g. based on network related distance information
between potential peers for P2P mode.
The CAN manager has interfaces with elements into the
Service Environment (HB or SP part - respectively CAN-HB
I/F and CAN-SP I/F), intra-domain network resource
manager and other domains’ CAN managers. The horizontal
interfaces allow negotiation between domains, based on
SLAs and supporting in this way synergic CAN functioning
over multiple domains (e.g. QoS enabled and controlled
paths crossing several CANs). Each CAN Manager
advertises its CANs possibilities to outside world and then
negotiates and concludes some SLAs with other domains.
D. Functional Aspects
While offering significant adavantages versus conventional
routers the CAN /MANE approach poses several open issues
and challenging research aspects. We briefly mention part of
them.
• CAN-related metadata processing: given the task of
performing content awareness, the CAN router should
use a flexible structured query mechanisms based on
pattern matching (pattern matching tools are needed to
parse the packet payloads and recognize predefined
sets of patterns, associated with algorithms for
different pattern matching data mining, and content
inspection subsystems).
• Protocol classification and description: CAN
subsystems should analyze the application and protocol
to classify and recognize traffic types. For this purpose
appropriate languages should be used for protocol
description and packet classification. They should
provide capabilities for customization and should be
flexible to be adapted to new applications and
protocols. Techniques for application packet analysis
have to be selected by comparatively analyzing the
•
•
•
•
•
•
approaches: stateful/stateless, signature based and
behavioral. Investigation of solutions is necessary to
offer best trade-offs between a high speed processing
requirement but also flexibility and low cost.
Content-based routing: the CAN router will realizes
decoupling between routing based on content
(identifiers, scope, service profiles, policies) and
forwarding, facilitating new communication paradigms
as anycast paradigms and publish-subscribe.
Content-based QoS: the CAN approach can allow to
the Network Provider to offer to the upper layers
specialised CANs optimised for different QoS classes
with configurable guarantees.
Content caching and processing in the network: CAN
concepts may help to better traffic engineering; the
current technologies make realizable storage functions
for large amount of data in the network nodes. Traffic
can be cached (when appropriate) in specially selected
nodes aiming to reduce/balance the network load,
reduce E2E delay, avoid nodes of the network to enter
congestion status. The real-time streams distributions
can be enhanced.
CAN layer monitoring: the CAN nodes current status
should be monitored from load point of view. The
CAN Manager receives performance information from
the network layer, and aggregates them in order to
evaluate path characteristics between CAN nodes
(distance).
Improved network security by using packet content
analysis: the packet content inspection can enhance the
system capabilities to detect intrusions and anomalies
which can jeopardize the network. Given that such
mechanisms are to be located in the network nodes,
high speed methods should be found.
CAN Performance aspects: CAN applies intensive
packet processing executed at network level (contrary
to the traditional IP network level), involving a
significant processing power. Therefore the challenge
in evaluating the performance versus complexity and
cost will be an important task of the ALICANTE
project. Cross-layer optimization capabilities are
possible at CAN layer interface with HB layer, as a
powerful tool offered by the CAN and network aware
applications approach.
E. End to end QoS issues
While CAN approach may offer a larger set of features,
the end to end (E2E) QoS assurance with different degrees of
guarantees seems to be natural one. SP can offer QoS
guaranteed services, realized at CAN level, by constructing
appropriate virtual (unicast or multicast) single or multipledomain pipes in the network, with or without resource
reservations, based on SLA contracts. Then, as a second level
of actions, adaptation actions will be performed as follows:
adapting flows proactively if we have , e.g. Scalable Video
Codec sources and while knowing that there are not enough
resources but still want transmission; or, dynamically in
reactive manner, done later if necessary, because of network
or terminal conditions variations. The services will be
accessible for users in two ways. The first is based on service
subscriptions done by the user via its HB (admission control
is applied by SP, in cooperation with NP, depending on
available resources) and then future invocations – with
guaranteed QoS levels based on resource reservations. The
second way is completely dynamic, based on invocation only
– but without having guaranteed chance of the call admission
success if guaranteed QoS is wanted. A new feature of the
proposed architecture is that it can offer a solution to better
QoS while working in P2P style at service level. An
appropriate signaling system should be developed in the
management and control plane to support QoS oriented
CANs.
V.
CONCLUSIONS
A novel layered architecture is proposed based on
Content Aware Networking (CAN) and Network Aware
Applications approach, with focus on the functionalities of
the virtual CAN layer. The work is a part of the starting
effort inside of a new European research FP7 IP project,
ALICANTE While the architecture fulfills the new
requirements for multimedia distribution and services via
multiple IP domains it is more general and tries to meet the
needs of the Future Internet. A flexible business model
supported by the architecture is introduced. The role of the
CAN layer, its requirements and interfaces with other system
layers are briefly described. QoS capabilities of the
architecture are summarized. Challenges are mentioned and
open research issues especially related to the
performance/cost of CAN devices in real network
environment. Significant future work will follow inside the
ALICANTE project for detailed specification, validation and
finally implementation on a large testbed.
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
ACKNOWLEDGMENT
[17]
This work has been partially supported by the European
Research Project FP7 ALICANTE No248652.
[18]
REFERENCES
[1]
[2]
T. Anderson et al., “Overcoming the Internet Impasse through
Virtualization,” Computer, vol. 38, no. 4, 2005, pp. 34–41.
J. Turner and D. Taylor, “Diversifying the Internet,” Proc.
GLOBECOM ‘05, vol. 2, 2005.
[19]
J. Schönwälder, et al., Future Internet = Content + Services +
Management, IEEE Communications Magazine, July 2009, pp.27-33.
Networked European Software and Sevices Initiative (NESSI)
Strategic Research Agenda, Vol. 3.FP7-2.exec, NESSI Roadmap,
Feburary, 2008.
N. M. Chowdhury and Raouf Boutaba, Network Virtualization: State
of the Art and Research Challenges, IEEE Communications Magazine,
July 2009, pp. 20-26.
European Commission, FP7 ICT Work Programme 2009-2010.
T. Kourlas, The Evolution of Networks beyond IP,
htp://www.iec.org/online/tutorials/acrobat/evolution_of_networks_be
yond_ip.pdf
Maria G. Martini, et.al., Content Adaptive Network Aware Joint
Optimization of Wireless Video Transmission, IEEE Communications
Magazine, January 2007, pp. 84-90.
C.Baladrón, User-Centric Future Internet and Telecommunication
Services, pp.217-226 (in G. Tselentis et al. (Eds.), Towards the
Future Internet, IOS Press, 2009)
T.Zahariadis, et. al., Content Adaptation Issues in the Future Internet,
pp.283-292 (in G. Tselentis et al. (Eds.), Towards the Future Internet,
IOS Press, 2009)
F.Liberal et al., QoE and *-awareness in the Future Internet, 293-302,
(in G. Tselentis et al. (Eds.), Towards the Future Internet, IOS Press,
2009)
N.Baker, Context-Aware Systems and Implications for Future
Internet, (in G. Tselentis et al. (Eds.), Towards the Future Internet,
IOS Press, 2009)
Amr El-Sherif, et.al., Content-Aware Cooperative Multiple Access
Protocol for Packet Speech Communications, IEEE Transactions on
Wireless Communications, Feb. 2009, pp.995-1005
E. Rainge, The Inevitable Failure of Content-Aware/DPI Network
Devices and How to Mitigate the Risk, September 2008 (adapted
from Worldwide Network Test and Measurement 2008.2012 Forecast
and 2007 Market Shares,
http://www.breakingpointsystems.com/resources/white-papers/idcwhite-paper/content-aware-testing.pdf
Á.Huszák, S.Imre , Content-aware Interface Selection Method for
Multi-Path Video Streaming in Best-effort Networks, http://www.ictoptimix.eu/images/6/61/ICT2009published.pdf
S.B. Kodeswaran, et.al., Content and Context Aware Networking
Using Semantic Tagging Data Engineering Workshops, 2006.
Proceedings. 22nd International Conference on, 2006, pp.67 - 77
V.Aggarwal, A. Feldmann, Can ISPs and P2P Users Cooperate for
Improved
Performance?,
ACM
SIGCOMM
Computer
Communication Review 31 Volume 37, Number 3, July 2007.
Xie, H., Krishnamurthy, A., Silberschatz, A., and Y. Yang, "P4P:
Explicit Communications for Cooperative Control Between P2P and
Network
Providers",
http://www.dcia.info/documents/P4P
Overview.pdf.
FP7 IP project, “MediA Ecosystem Deployment Through Ubiquitous
Content-Aware Network Environments”, ALICANTE, No248652,