World Wide Wireless Web.(wwww)-Part4

ign="center">various network reoptimisaiton techniques in traffic
World Wide Wireless Web.-Part 4. Integration ofengineered MPLS networks. Dr. Jay Perrett is Chief
MPLS with 3G/4G, WLAN and other wirelessScience Officer & Head of R&D with Aria Networks
technologies.Limited where he works on the development of
BELOW ARE SOME BEGINNINGS OF FORAY INTOadvanced machine learning solutions to complex path
COMPLETE WIRELESS INTEGERATIONcomputation and network reoptimization problems.
Scaling MPLS in Multi-Area Networks Bruno Decraene,Before joining Aria Networks, Dr. Perrett was CTO of
France Telecom Next generation networks use andApplied Insilico and applied machine learning techniques
provide VPN services and therefore need MPLS. Theto the drug discovery and biological research industries.
size of these networks is big and is getting bigger asWith a long history in Academia and the
IP/MPLS is spreading in aggregation networks, sotelecommunications industry, spanning UK and US
several IGP areas are common. However extendingUniversities and Research institutes and companies
LSPs across areas is quite a challenging task. Thisincluding Lucent Technologies, Redback and Movaz
presentation covers the set up of MPLS LSPsNetworks, Dr. Perrett is ideally placed to bring together
between nodes belonging to different IGP areas, whenthe advanced problem-solving abilities of
Traffic Engineering is not required. We first present anext-generation machine learning applications with the
case based on current service provider requirements.high demands of path computation in complex and
Then, we describe existing mechanisms to addressgrowing converged Next Generation Networks. Open
this case and discuss their limitations . We next presentIssues in Hierarchical Recovery Payam Torab, Lambda
a new solution using an LDP extension to ease theOptical Systems Generalized Multiprotocol Label
set-up of inter area LSPs and better address currentSwitching (GMPLS) is recognized as a universal control
networking issues. Finally we provide a high levelplane with wide applicability to various switching
comparison of the existing and new mechanisms intechnologies, ranging from packet switching in IP
terms of performances, impact on IP routing,networks to photonic wavelength and waveband
survivability, and scalability. Verizon Business MPLSswitching in optical WDM networks. Today, this wide
Enabled Core Network Design Studies: using WANDLapplicability is being demonstrated by the emergence
IP/MPLS View Ning So, Verizon Business In today'sof multi-region networks, where a single instance of
economic environment, it is uncommon for a carrier toGMPLS control plane controls multiple switching
build a core network from scratch. Instead, thetechnologies (regions) at the same time. In this
establishment of any core network is often the resultpresentation, we examine a few areas around fault
of the natural evolution of an existing backbonemanagement and recovery in multi-region networks. In
network. Although this method often result in anparticular, we study 1) GMPLS applicability to
extended network build out, and makes the networkend-to-end recovery in these networks, and potential
design more challenging, the benefits are readilyareas for protocol extension, and 2) the implications of
apparent. Networks can realize immediate relife fromoptical transparency on fault management, in particular
bottlenecks. Also network's return on investment (ROI)fault localization, in case one or more network regions
will accelerate as new services are introduced due toemploy all-optical switching. Regarding GMPLS
the ability to use the latest technologies. This jointapplicability to recovery, we look at the standard
presentation from Verizon Business and WANDL, Inc.GMPLS end-to-end recovery techniques available
will address several network design challenges thattoday, and how they can be used to achieve
Verizon Business faced when designing and building itsend-to-end recovery in multi-region networks. We will
MPLS-enabled Core Network. Each challenge will bediscuss the following points and more: . Since
discusses with examples, and the solutions will beend-to-end recovery options in one region correspond
presented. Verizon Business Design Challenge #1:to link recovery options in another region, there has to
Designing an MPLS core network that can sustain dualbe a one-to-one correspondence between end-to-end
network failure - As the trend of networkand link recovery options in GMPLS, which is not the
convergence onto a single MPLS core backbonecase at this time. In particular, the notion of link
continues to grown, and the transmission technologyrestoration is currently missing from the GMPLS
moves from the conventional ring based SONETstandard. . Assuming the GMPLS end-to-end and link
architecture into linear Ultra Long Haul technology,recovery options are unified, we discuss the potential
designing an MPLS core network that has the ability toneed to signal or distribute additional information to
restore 100% of traffic during dual network elementcoordinate recovery at different regions. In particular,
failure becomes critical. Verizon Business Designrecovering an H-LSP can affect the diversity of paths
Challenge #2: Designing an MPLS network withthat use the H-LSP as a link, a fact that has not been
backbone trunks of various sizes - Carrier backbonestudied with sufficient detail so far. A related problem
networks have traditionally used a tiered structure withto recovery in multi-region networks is the implications
one backbone size for the core backbone, andof optical transparency on fault management.
another bandwith size for edge-to-edge, edge-to-coreMulti-region networks with an optical core are moving
trunks. In MPLS based networks, backbone trunkstowards transparent optical switching for scalability and
could range from OC12 all the way up to OC768. Thiseconomic reasons, but this also makes SONET-style
presents a new set of MPLS traffic engineeringperformance monitoring and fault management at
challenges, especially during the network failures.each node more difficult. Network performance
Verizon Business Design Challenge #3: Designing anmonitoring can still be done at each node, but often
MPLS network with ever more sophisticated customerwith limitations. Of particular interest is the complexity
routing requirements - Verizon Business is experiencingof localizing soft failures such as BER degradation, and
ever more sophisiticated customer routinghow one can benefit from the vertical hierarchy of
requirements for its traffic. MPLS-based networkmulti-region networks to make fault localization faster
makes it possible for carriers to satisfy theirand more efficient. We share lessons from our
customer's unique requirements; it also createsexperience with a multi-region network that does not
network design challenges. The design of the MPLShave one, but two layers of optical transparency: A
core network is an evolving process. WhileDWDM network with multi-granular optical
MPLS-based networks introduce many advantages,crossconnects (MG-OXCs) that support transparent
their inherent flexibility also introduces many challengesswitching at both waveband and wavelength levels.
which need to be addressed during design process. IsHere, we show that the GMPLS fault localization
your MPLS Network Ready for IPTV? Martin Lai,performance (in terms of speed and usage of
Agilent Technologies The introduction of Triple Playmonitoring resources) for soft failures can be
services such as IPTV represents a new and excitingsignificantly improved if the vertical hierarchy of
time for consumers. The expected financial rewardsmulti-region networks is considered in the design. We
from IPTV and other multicast services are triggering aconclude the presentation with final thoughts on
remarkable evolution in Telco networks this decade.applicability of GMPLS as a fast, reliable and robust
Network equipment manufacturers and servicecontrol plane technology for multi-region networks.
providers are now faced with the challenge ofFuture Directions in MPLS QoS Bruce Davie, Cisco
ensuring their devices and networks can meetSystems Since the early days of the MPLS there
scalability, performance and reliability expectations forhave been great expectations that it could improve the
multicast services including IPTV. There are severalQoS capabilities of IP networks. In reality much of the
options being proposed within the IETF to ensureMPLS QoS architecture simply mimics the capabilities
efficient multicast service delivery over MPLSthat have been developed for IP, such as Diffserv.
networks. In this presentation we will discuss severalHowever, MPLS has had a significant impact on QoS
of the mechanisms being proposed includingfor a variety of reasons. MPLS VPNs have turned out
Point-to-Multipoint (P2MP) extensions, multicast VPNsto be the catalyst for a great deal of Diffserv
(mVPNs) and multicast VPLS. We will describe thedeployment - indeed the majority of service provider
test methodologies involved in validating theDiffserv deployments are in the context of MPLS
interoperability, scalability and performance of devicesVPNs. And the decision to use RSVP as the signaling
and networks implementing these approaches. You willprotocol for MPLS-TE has provided a key building
also be presented with key test scenarios to evaluateblock for scalable admission control - an extension to
and compare the different mechanisms available forthe IP architecture that seems likely to be increasingly
efficient multicast delivery and measure the impactimportant with the advent of large scale video over IP.
they have on the end user viewing experience.This talk will explore the future direction of MPLS QoS,
Enabled Operational Modes Using PCE-Basedincluding the development of more scalable admission
Architecture Emmanuel Dotaro, Alcatel The emergingcontrol techniques based on congestion marking, and
PCE protocols open the door for valuable newthe challenges of delivering end-to-end SLAs over the
operational modes. The PCE is for example annetworks of multiple independent providers.
efficient enabler for vertical integration (betweenConclusion: The 3G cellular networks, e.g. UMTS are
layers) and horizontal integration between domains.designed to provide users with voice and data
This presentation explore some innovative modesservices. Total cell capacity limits the per user data
such as Interdomain TE scenarios, enhanced overlayrate. Many times highspeed requirements are clustered
models using PCE-to-PCE enriched communications,in small pockets. These clusters are termed as
provisioning and recovery performances improvments.hot-spots. Network operators would like to employ
Beyond the protocols mechanics, the presentationefficient solutions, which are easily integrated with their
aims at demonstrating the potential key role of theexisting UMTS, based infrastructure.WLANs offer an
PCE-based architecture which may foster undeployedattractive solution. 3G cellular accesses based on code
technologies existing but sometimes considered toodivision multiple access (CDMA) either wideband
complex. The combination with a Policy BasedCDMA or cdma2000 can be used to satisfy users
Management architecture is thus falling into the scopewho have a larger need for mobility while 802.11
of this introduction to PCE-based architecturesystems can be used to support users with much
applicability. Increasingly Complex Demands for Pathlesser coverage area requirements. It is in light of this,
Computation in Converged Networks Jay Perrett, Ariathe next wave of technological advance is already
Networks The trend for convergence of triple playunder consideration i.e. 4G. Several International
services over IP/MPLS networks places increasedprojects are already underway to ensure that
demand for complex service placement algorithmsservices are suited to the characteristics of several
within the network. It is no longer sufficient to throwdifferent delivery mechanisms, from cable networks to
bandwidth at the problem as Service Providers mustGPRS networks, and the services can be delivered
extract better value from their deployed resourcesusing the most appropriate of the range of networks
and must distinguish the traffic for different servicesavailable. In addition the dynamic allocation of available
they offer. High specification SLAs force networkspectrum between different wireless networks is also
operators to think carefully about the bandwidth andunder investigation. The challenge is to explore the
quality guarantees that they are making, and requiredesign of such a transport infrastructure which will be
sophisticated network planning to ensure that a varietyable to take full advantage of IP based technologies
of conflicting constraints and demands are met forachieving desired mobility between the various access
each customer service. This presentation examinestechniques and at the same time provide the
how the requirements of voice, video and data trafficnecessary capabilities in terms of QoS, robustness and
map to different path computation requirements formanageability. The goals at the present stage
minimized cost and delay, maximized resilience, andregarding the development of mobile standards
guaranteed bandwidth. The different objectives ofremains common (3GPP and 3GPP2) and include IP
Fast Reroute and end-to-end protection will bebased multimedia services, IP based transport and the
discussed and the consequences for network planningintegration of IETF protocols for functions such as
and path computation will be highlighted. The speakerwide area mobility support (MIP) , signaling (SIP) and
will identify how different computation techniques canauthentication, authorization and accounting (AAA), it is
achieve suitable paths only under specific conditions,popular to call any network that satisfies these criteria
and will go on to examine the benefits and risks ofas an all-IP network.