DOCX by changcheng2

VIEWS: 14 PAGES: 10

									                                            Future Internet
                                 Department of Advanced Information Technology
                                         Professor OKAMURA’s Laboratory.
                          Othman Othman M.M., Master course student, 1st year.
                    Proceeding the First Exercises on Computer and Systems Engineering.

                                                   January 12, 2010.


                                                               about the future internet apart from technology and
1. Introduction:                                               specific details. The second paper is the A”KARI
                                                               Conceptual Design paper”, which provides more specific
Since the beginning of the Internet 40 years ago it played     details about needs required form the future internet
a great role in the development of mankind. But as time        and their correlating design requirements, and the
passes human needs become more demanding and more              design principles according to the AKARI project. While
complicated. The thing that makes the Internet which           the third paper which is the “GENI overview”, describes
was designed for simple reliable communication of              the creation of GENI that is a testbed network for the
packets fails to cope with the new needs of the future.        future internet research. And the Last paper to be
For that reason the importance of the future internet          discussed is the OpenFlow which is considered to be one
rises to overcome the limitations of the current internet      of the promising technologies enabling the future
and to provide more functionalities that the current           internet.
internet failed o provide. As an example in current
internet bottlenecks usually occur at the server side.
Although many methods were proposed to solve this
problem but they still work under the same limitations of
the current internet. And so proposing a new method
                                                               2. EIFFEL report:
that makes use of the rapidly increasing internet speed
of the home users for solving this problem will                2.1. Brief introduction
contribute to develop internet, take some burden off the
server side and helps to balance the traffic.                  On the European side of the future internet design the
                                                               7th framework program’s EIFFEL think tank was launched
In a bit more detail, currently servers are facing an          for the purpose of developing the future internet. In this
increasing demand causing the bottleneck to be on the          report future internet is studied among many axes
server side. Many ways are for solving these problems
                                                               Technology, economics and society.
were proposed like increasing the capabilities of the
server side (bandwidth and power of server), the content
delivery networks, any cast and many others. But those         2.2. Technology Axis
methods had some limitations like the lack of dynamicity       First of all, on the technology axis. It is difficult to study
and the redundancy. But currently the bandwidth of             about future internet since it is difficult to see or expect
home users is increasing dramatically, and so a new
                                                               the future. And so looking backward to see the reasons
method that makes use of this increased bandwidth to
help in increasing the number of clients getting the           behind the current internet’s success would be beneficial.
service.                                                       It is believed that the original design of the current
                                                               internet architecture lies behind its success. Another
In this lecture many papers related to the future internet     reason behind the current internet success is the
are discussed. The first one will be the “EIFFEL report”
                                                               applications and the variety of types (e.g: eBay, Amazon,
which provides a kind of brain storming and many ideas
                                                               Google, Skype, BitTorrent …). Those successful internet
and lessons for the future internet along tree axes
technology, economics and society. It provides rich ideas      applications are invented by practitioners because they
                                                               understand their needs and they invent solutions to suit

                                                                                                                          1
them. It is also important to observe factors that allowed    value chain, a certain degree of sustainability is required
and enabled the development of the internet that are          to make investment happen and maintain a health value
the enablers for the user community which are: the            chain. In order to maintain healthy value chains it is
abundant and affordable computing resources provided          important to be able to account for the usage of
by Moore’s law and the universal connectivity provided        resources along the value chain. This requires
by the internet.                                              mechanisms and overall architectural approaches for a
                                                              resource accountability framework. Third is the current
In the study of designing the future internet it is           business structures and their validity to the future. Any
important to understand that the both engineering and         fundamental approach to change the current Internet
science must cooperate. Engineering focuses on                will undoubtedly have an impact on the existing
designing the best networks. While science focuses on         business structures.
studying the nature of networks. And so the design
process should start with the engineering design of the       2.4 Society Axis
system and then studying it scientifically. By doing so,
the finding of the scientific phase can be feed back to the   Finally, on the society axis. Internet reflects parts of the
design or revision phase.                                     social structure and it has a strong influence on the social
                                                              structure. Also the anti-social activates do exist like the
Finally lessons should be taken from the current internet.    attacks and so on. Also it is important to notice that
First one is the new demands on the architecture caused       enovation will always have element of surprise on how
by the advancement of technology and internet; those          society will coop with this enovation and how it will
demands should be continuously identified and                 make use of it or evolve it.
addressed. The second one is the big gap between the
reality and the research community understanding of it,       Another social aspect is the importance of studying the
which should be reduced. And the last one is the              information overload, since on the internet there is too
understanding that successful internet architectures          much information, too many services, too many options
should evolve by time.                                        and too many individuals. And so focusing on be to
                                                              evolve approaches that reflect the human and social
2.3. Economics Axis                                           approaches to dealing with overload.


Second, on the economics axis. As the internet started it     Another aspect to be studied is impact of governance on
was designed to place ends in the center and assuming         internet and the effect of internet on governance.
that we have dumb network. This led to enovation on
the ends creating services like Google and many others,
                                                              3. AKARI Conceptual Design:
but this left many problems unsolved like QoS,
multicasting… . Addressing those problems will require        3.1 Brief introduction
economic coordination among many stakeholders. Also
special care for considerations like deploy ability and       AKARI which means “a small light” in Japanese, is the
economic sustainability is crucial. All those                 Architecture Design Project of the Network Architecture
considerations require the cooperation of the technology      Group of the New Generation Network R&D in the NICT
and the economics communities.                                (National Institute of Information and Communication
                                                              Technology).
Some of the challenges that require joint work will be
described here. First is the joint system design, it is
important to accommodate the differences of research
styles that exist between research fields like economics,
engineering and social science. Second is the sustainable

                                                                                                                        2
3.2 Goals of AKARI project                                     3.4 Societal considerations and their correlating design
                                                               requirements
The AKARI project’s main goal is to design the network of
the future which will serve the human development for          The internet is no longer a mean of communication; it is
50 or 100 years. AKARI aims to do so by building               currently playing a role in formulating the society. And
technologies for the new generation network by 2015,           according to this many societal considerations must be
developing network architecture and creating a network         taken into account like:
design based on that architecture.                               1. Peat-bps class backbone network, 10Gbps FTTH, e-
                                                                     Science. fig.3 shows the grow of traffic while fig.4
                                                                     gives an example of the use of internet and its
3.3 Why to use the clan state principle                              frequency of use.
                                                                 2. 100 billion devices, machine to machine (M2M), 1
The AKARI project follows the clean state philosophy of
                                                                     million broadcasting stations.
design which means without being restrained with the
                                                                 3. Principles of competition and user-orientation.
current network technologies and without being
                                                                 4. Essential services (medical care, transportation,
impeded to the current constrains. The clean state idea
                                                                     emergency services), 99.99% reliability.
originated to overcome difficulties of the current
                                                                 5. Safety, peace of mind (privacy, monetary and
internet, since functions and layers were added on top of
                                                                     credit services, food supply traceability, disaster
each other for 30 years as seen in fig.1 making it difficult
                                                                     services).
to ensure reliability of the current internet which
                                                                 6. Affluent society, disabled persons, aged society,
became complex and incompatible. For example, current
                                                                     long-tail applications.
internet suffers from loss of transparency caused by the
                                                                 7. Monitoring of global environment and human
wide spread of Network Address Translation (NAT) used
                                                                     society.
due to the insufficient number of IP addresses. Another
                                                                 8. Integration of communication and broadcasting,
example is the slow recovery from failures in the Inter-
                                                                     Web 2.0.
domain routing table. With the large number of those
                                                                 9. Economic incentives (business-cost models).
limitations the clean state offers a new beginning for the
                                                                 10. Ecology and sustainable society.
future internet to avoid such limitations.
                                                                 11. Human potential, universal communication.




fig.1: Rapid adding of functions and layers in the current
                         internet.                                             fig.2: Capacity vs. Access frequency.




                                                                                                                          3
                                                                   the global environment from various viewpoints is
                                                                   indispensable for accomplishing this.
                                                               8. Integration and simplification. The design must
                                                                   be simplified by integrating selected common
                                                                   parts, not by just packing together an
                                                                   assortment of various functions. Simplification
                                                                   increases        reliability     and      facilitates
                                                                   subsequent extensions.
                                                               9. Network model. To enable the information network
                                                                   to continue to be a foundation of society, the
                fig.3: Grow of internet traffic.                   network architecture must have a design that
                                                                   includes a business-cost        model    so      that
And to deal with those considerations the following                appropriate economic incentives can be offered
design requirements must be met:                                   to     service providers and businesses in the
 1. Large capacity. Increased speed and capacity are               communications industry.
     required to satisfy future traffic needs, which are       10. Electric power conservation.          As network
     estimated to be approximately 1000 times current              performance increases, its power consumption
     requirements in 13 years as shown in fig.3.                   continues to grow, and as things stand now, a
 2. Scalability. The devices that are connected to                 router will require the electrical power of a small-
     the network will be extremely diverse, ranging                scale power plant. The information-networked
     from high-performance servers to single-function              society of the future must be more Earth friendly.
     sensors. Although little traffic is generated by a        11. Extendibility. The network must be sustainable. In
     small device, their number will be enormous,                  other words, it must have enough flexibility to
     and this will affect the number of addresses and              enable the network to be extended as society
     states in the network. This can be shown in fig.2             develops.
     which shows different internet uses along with its
     frequency of use and its bandwidth demand.
 3. Openness. The network must be open and able to
     support appropriate principles of competition.           3.5 Basic design principles
 4. Robustness. High availability is crucial because
     the network is relied on for important services
     such as medical care, traffic light control and
     other vehicle traffic services, and bulletins during
     emergencies.
 5. Safety. The architecture must be able to
     authenticate all wired and wireless connections.
     It also must be designed so that it can exhibit safety
     and robustness according to its conditions during a
     disaster.
 6. Diversity. The network must be designed and                               fig.4: Basic design principles.
     evaluated based on diverse communication
     requirements      without      assuming       specific    1. KISS (Keep It Simple, Stupid): important for
     applications or usage trends. As shown in fig.2.             increasing the diversity, expandability and reliability.
 7. Ubiquity. To implement pervasive development                  And to support that :
     worldwide, a recycling-oriented society must be
     built. A network for comprehensively monitoring


                                                                                                                       4
   a. Crystal Synthesis: simplification of technologies
      to reduce complexity even when integrating
      functions.                                                  b. Bi-directional authentication:
   b. Common Layer: new generation network                        c. Traceability: to reduce attacks on the network.
      architecture will have a common layer and will                 Important to consider in designing addressing,
      eliminate redundant functions in other layers to               routing and transport on top of them.
      degenerate functions in multiple layers. To
      illustrate the common layer concept, fig.5 shows        3. Sustainable and Evolutionary principle: New
      the current internet’s common layer.                       generation network must be sustainable network
                                                                 that can evolve and develop in response to
                                                                 changing requirements.
                                                                 a. Self-*     properties:     self-distributed, self-
                                                                    organizing, self-emergent.
                                                                 b. Scalable, distributed controls: To scale controls
                                                                    even in large scale or topologically varying
                                                                    networks,
                                                                 c. Robust large-scale network: As the scale or
                                                                    complexity increases the rate at which errors
           fig.5: Common Layer in the current internet.             occur increases. And so the new generation
                                                                    network must be designed to handle
   c. End to End: network should not be constructed                 simultaneous or serious failures that may occur.
      based on a specific application or with the                d. Controls for a topologically fluctuating
      support of a specific application as its                      network: in P2P or mobile network topology
      objective.                                                    frequently changes controls for finding
                                                                    resources on demand are more effective than
2. Reality Connection principle: network entities must              controls for maintaining routes or addresses.
   have association with real world society. To support             Balance must be made due to high overhead of
   this association:                                                finding resources on demand.
   a. Separation of physical and logical addressing: to          e. Controls     based     on      real-time    traffic
       support mobility and multi-homing. Fig.6 shows               measurement: routing must adopt controls of
       an example of splitting the identifier and locator.          precision-optimised         real-time       traffic
                                                                    measurement.
                                                                 f. Openness: openness to users encourages
                                                                    innovation.


                                                             5. GENI System overview:
                                                             4.1 Brief introduction

                                                             GENI which stands for “Global Environment for Network
                                                             Innovations” is a suite of network research infrastructure
                                                             funded by the National Science Foundation (NSF) to
                                                             support experimental research in network science and
                                                             engineering. The intellectual space of GENI ranges from
            fig.6: Separation of Locator and Identifier.     new research in network and distributed system design

                                                                                                                     5
to the theoretical underpinnings of network science,
network policy and economics, societal values, and the
dynamic interactions of the physical and social spheres     4.3 Design Goals
with communications networks.
                                                            And to ensure the resulting infrastructure suite will be
                                                            useful to the research community the following design
4.2 Core Concepts
                                                            goals are used to design GENI:
                                                             1. Generality: each experimenter has the flexibility to
The core concepts of GENI gives a boarder view
                                                                 perform the desired experiment. This means that
understanding of GENI:
                                                                 each component should be programmable.
 1. Programmability – researchers may download
                                                             2. Diversity & Extensibility: GENI must include a wide
    software into GENI-compatible nodes to control
                                                                 class of networking technologies, spanning the
    how those nodes behave. As it is shown in fig.7.
                                                                 spectrum of wired and wireless technologies
 2. Virtualization and Other Forms of Resource Sharing
                                                                 available today.
    –nodes implement virtual machines, which allow
                                                             3. Fidelity: ability to support experiments that
    multiple researchers to simultaneously share the
                                                                 correlate to what one might expect in a real
    infrastructure. fig.7 shows the use of virtual
                                                                 network.
    machines.
                                                             4. Observability: support for measurement-based
 3. Federation – different parts of the GENI suite are
                                                                 quantitative research.
    owned and/or operated by different organizations.
                                                             5. Ease of Use: researchers must able to make full use
 4. Slice-based Experimentation – GENI experiments
                                                                 of its federated infrastructure.
    will be an interconnected set of reserved resources
                                                             6. Sliceability: GENI must be a shared infrastructure
    on platforms in diverse locations. Researchers will
                                                                 suite in order to achieve cost effectiveness.
    remotely discover, reserve, configure, program,
                                                             7. Controlled Isolation: must support strong isolation
    debug, operate, manage, and teardown distributed
                                                                 between slices.
    systems established across parts of the GENI suite.
                                                             8. Opt-in: must make it easy for a broad mix of users
    fig.7 illustrates a slice for running the researchers
                                                                 to “opt in’’ to experimental services. Which means
    software.
                                                                 to give board range of users the ability to
                                                                 participate in the experimental service.
                                                             9. Security: must be secure, so that it cannot
                                                                 accidently or maliciously be used to attack today’s
                                                                 Internet.
                                                             10. Federation & Sustainability: must be designed for a
                                                                 15-20 year lifetime.


                                                            4.4 Building Blocks

                                                            GENI system can be broken down into its building blocks
                                                            which are:


           fig.7: Illustration of the core concepts.




                                                                                                                  6
                                                           fig.8: Building blocks and slice.

                                                                       1. Resource discovery: in this stage the researcher
 1. The researcher: is the one who wants to run an                        finds out about the available components through
    experiment.                                                           his federation’s clearinghouse. See fig.9.
 2. Components: is the primary building block and it                   2. Slice Creation: in this stage the resources within this
    encapsulates a collection or resources.                               slice are linked together to form a coherent virtual
 3. Aggregate: is an object representing a group of                       network in which an experiment can run. This stage
    components.                                                           is also done through the clearinghouse. See fig.10.
 4. Clearinghouse: it is a group of authorities like the               3. Experimentation: in this stage the researcher can
    trust, slice and services. It performs tasks like                     download code into his slice, debug, collect
    managing trust between different federations,                         measurements, and iterate. This stage is conducted
    resource discovery, slice creation and many others.                   directly between the researcher and the
 5. GENI Operations and Management: is a system-                          aggregates. . See fig.11.
    wide function that keeps GENI resources operating                  4. Modifying a Slice: while running experiments
    and manages GENI services. And it plays roles like                    researcher can modify his slice by adding or
    stopping an overrun experiment or malicious                           removing some components to his existing slice.
    experiments.                                                          This stage is some kind similar to the resource
 6. Federation: federation permits the interconnection                    discovery and the slice creation stages, and it is
    of independently owned and autonomously                               done also through the clearinghouse. . See fig.12.
    administered facilities.                                           5. Stopping the experiment: stopping of the
 7. Slices: is a substrate-wide network of computing                      experiment can be done by the researcher after
    and communication resources.                                          finishing his experiment or might be done by the
                                                                          GENI operations and management as an emergency
4.5 Stages of experiment                                                  procedure to overcome an overrun or malicious
                                                                          experiment. . See fig.13.
And in order for a researcher to conduct an experiment
he has to go through one or more stages listed here:




                                                                                                                               7
fig.9: Resource Discovery.             fig.12: Expanding a Slice and Federations cooperation.




                                                   fig.13: Stopping an experiment.

   fig.10: Slice Creation.


                              5. OpenFlow:
                              5.1 Brief introduction

                              OpenFlow was created in Stanford University as part of
                              their clean slate project that researches future internet.
                              In the current internet researchers are conducting
                              experiments on test beds or small environments that
                              they create, but they did not have the ability to run
                              experiments on the real world’s networks. That is
                              because of the enormous number of different protocols
                              and different vendor’s equipments installed, it is also
   fig.11: Experimentation.   feared that their experiments might disrupt the
                              production traffic (normal traffic) and because the
                              network equipment vendors are not willing to open their
                              equipment to be modified by researchers. The thing that
                              created a barrier for researcher’s innovation. That was


                                                                                                8
where the idea of OpenFlow first arise, to create an open,
programmable, virtualized platform to run the
researchers experiments on their campus network
without affecting the production traffic.



5.2 How Openflow works

To shade more light on OpenFlow, it is important to
show its main components and to show how it works.
fig.14 shows the components of OpenFlow system. The
first component of the OpenFlow system is the
OpenFlow switch or router, which has the regular routing
or forwarding table and a new table called the flow-table
which is similar to the TCAM (Traffic Control and Monitoring
used for firewall and other purposes). The OpenFlow switch            fig.14: Overview of OpenFlow system.
uses the flow-table to decide how to forward the
experimenters data according to his settings and uses
the regular routing or forwarding table to process the
                                                                                      Ethernet               IP                    TCP




                                                                                                                                                   Statistics
                                                                       VLAN ID
production traffic. Also the OpenFlow switch uses the




                                                                                                                                          Action
                                                                In
OpenFlow Protocol which acts as a secure channel that          Port                SA SA Type     SA    SA        Protocol   Src    Dst

allow the controller to add or remove entries in its flow-
                                                                                                 fig.15: Header in flow-table.
table. The second component is the Controller, which is
responsible for adding and removing entries to the
OpenFlow switch’s flow-table through the OpenFlow               5.3 Example of using Openflow
protocol on behalf of the experimenters. The controller
might be the one making decisions and adding flows for          An example of the use of the OpenFlow in fig.16 is using
one experiment or might be supporting many                      it to implement a network-wide policy for a campus
experiments; also it might be controlling one switch or a       network in a central controller. This controller can apply
whole network of switches. Finally, the last part to shade      polices like allowing gests to use HTTP but through a
light on will be the flow-table. The flow table entry           special proxy, preventing VoIP phones from
consists of a header so that incoming packet’s header           communicating with laptops, and any other policy
can be compared to it and an action related to that             needed.
header. The header as shown in fig.15 can be one or
more of the in port, VLAN id, TCP source, TCP destination,
network protocol, IP source, IP destination, Data Link
protocol, MAC source, MAC destination. And the action
related to the header can be: forwarding the packet to a
specific port or ports, or encapsulating the incoming
packet header and sending it to the controller so the
controller will decide what to do and takes action by
adding flows, or dropping the packet for some purpose                            fig.16: Campus wide security policy using OpenFlow.
like security, or forwarding the packet according to the
switch’s normal switching table.




                                                                                                                                                                9
References:


[1] Masaki Hirabaru, Masugi Inoue, Hiroaki Harai, Toshio
Morioka,                                             Hideki
Otsuki, et al: “New Generation Network Architecture ,
AKARI                                           Conceptual
Design      (ver1.1)”,   October     2008,     http://akari-
project.nict.go.jp/eng/index2.htm.
[2] Dirk Trossen (ed), Bob Briscoe, Petri Mahonen, Karen
Sollins,                                                Lixia
Zhang, Paulo Mendes, Stephen Hailes, Borka Jerman-
Blaciz,                                              Dimitri
Papadimitrou : “EIFFEL Report: Starting the Discussion” ,
13                 TH               OF                 JULY
2009,                           http://www.fp7-eiffel.eu/.
[3] The GENI Project Office: “GENI System Overview” ,
Document                                                  ID:
GENI-SE-SY-SO-02.0,        September        29,       2008,
http://www.geni.net/.
[4] Nick McKeown, Tom Anderson, Hari Balakrishnan,
Guru                     Parulkar,                     Larry
Peterson, Jennifer Rexford, Scott Shenker, Jonathan
Turner:                                        “OpenFlow:
Enabling Innovation in Campus Networks”, ACM
SIGCOMM Computer Communication Review, Volume
38        ,        Issue      2         (April       2008).




                                                                10

								
To top