An Embedded Web Server Architecture for XML-Based Network Management - DOC by liy18007


									                                            UNIVERSITE LIBRE DE BRUXELLES

                                                FACULTE DES SCIENCES

                                       SERVICE TELEMATIQUE ET COMMUNICATION

        Management Solutions for the New Communications World
           Network Operations and Management Symposium
                              IEEE IFIP

                                           April 15th -19th, 2002, Florence

                                                   Nadia Paroni

Faculté des Sciences                                                          ULB - Rapport STC
Service Télématique et Communication                                          02-01
CP 230 - Blvd du Triomphe                                                     septembre 2002
B-1050 Bruxelles
                                         NOMS 2002

    Management Solutions for the New Communications World
        Network Operations and Management Symposium
                                              IEEE IFIP

                                         April 15-19, 2002
                                          Florence, Italy

NOMS 2002, held in Florence, Italy from 15 to 19 April, is the 8th Network Operations and
Management Symposium. It presents a series of technical conferences in the area of network
and systems management and operations, jointly sponsored by the IEEE Communications Society
and IFIP (International Federation of Information Processing) societies.
Following the symposium theme "Management Solutions for the new Communications World",
NOMS 2002 has presented the latest approaches and technical solutions in the area of network
operations and management. The stress has been put on the major management issues for the
emerging telecommunications environments, such as 3 rd Generation mobility, Telephony over IP,
Optical Networking, …
In the technical program, 24 technical sessions were scheduled, divided in three parallel tracks:
two focusing on various aspects of the network management field from a technical and academic
points of view, and a third track covering paper presentations and panel discussions on the
operational aspects, industry experiences and business aspects.
NOMS combined in-depth technical sessions and lively panels with a demo/exhibition area
showing the latest network and service management technologies and products.
Coming from the four corners of the planet and representing all different sectors of the
telecommunications industry, top-level keynote speakers have daily opened the technical sessions
with their forward-looking views, and a distinguished experts panel has provided the grand
closing session.
Here are summaries of some technical sessions, that we attended, covering a number of selected
topics. More information can be found in the Symposium proceedings.

Network Operations and Management Symposium                                        Page 1
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

     An Embedded Web Server Architecture for XML-Based Network
             Hong-Taek Ju, Mi-Jung Choi, Sehee Han, Yunjung Oh, Jeong-Hyuk Yoon,
                                Hyojin Lee and James W. Hong
              Dept. of Computer Science and Engineering, POSTECH, Pohang, Korea
Embedded Web servers are widely used today for IP-based element management. This
intervention presents a new management architecture combining this technology with XML 1,
DOM2, and XPath3 to unify element management and network management. By taking advantage
of modern Web technologies, the proposed architecture provides a method to develop
management applications efficiently and to manage network devices
A typical case of applying Web technology to network management is to embed a Web server
into a network device for element management. This type of Web server is called an Embedded
Web Server (EWS). A EWS provides users with a Web-based management interface
constructed using HTML, graphics, and other features common to Web browsers. The status of
a device is provided to the user by simply retrieving pages, and an operator command is sent
back to the device using forms that the user completes. Accessing Web-based management user
interfaces through embedded Web servers offers many advantages: ubiquity, user-friendliness,
low development cost, and high maintainability.
The Simple Network Management Protocol (SNMP) is the network management solution used by
most of the industry since the early 1990s. It provides an interoperability framework to collect
instrumentation-level measurement data to support fault and performance management.
Compared to SNMP-based element management, Web-based element management significantly
improves configuration management, remote diagnosis, and remote troubleshooting. Each
management scheme presents advantages and disadvantages. Consequently, the two coexist in
the real world.
The concurrent support of SNMP- and Web-based element management poses a problem,
because it requires more computing resources in each network device. If the same device
supports two agents, one for SNMP and another for HTTP, the management overhead can
become too high. The main motivation behind the new management architecture is to use
embedded Web servers for network management as well as element management.
Another motivation is to apply XML for encoding management data into the HTTP payload. The
third motivation is to use XML Schema for specifying management information. XML is more
than a document markup language: it also helps model content and create standards for content.
The objective of this intervention is to present some of current realisations; EWS has been
extended from element management to network management by using standard Web

  Extensible Markup Language
  Document Object Model
  expression language used to address parts of an XML document
Network Operations and Management Symposium                                       Page 2
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

technologies. The XML Schema has been leveraged for specifying management information and
use XPath for the addressing scheme.
The further work consists of
-   developing a global management system for a commercial Linux server based on the proposed
-   looking to integrate other organization models such as policy-driven management,
    management by delegation, mobile agents, and intelligent agents into the XNM architecture.

                Model-Driven Open Packet Telephony Management
                            A. Clemm, P. Leung Cisco Systems, Inc.
                     170 West Tasman Drive San Jose, CA 95134-1706 USA
Open packet telephony (OPT) networks that carry voice over IP (VoIP) or voice over ATM
(VoATM) are becoming increasingly popular among service providers. Along with large-scale
deployments, management solutions are starting to appear. By their nature, OPT networks
depend on the interplay of large numbers of network components of many different types, all of
which are independently managed. Consistent provisioning of those components, detection of
mismatches in environments not properly controlled through an Operations Support System
(OSS), and the correlation of alarms and impact analysis of component failures on services are
among the challenges that service providers face. Because of this, there is an increasing need
for truly end-to-end, comprehensive solutions for this type of technology.
The current state of the art is characterized by solutions that are fairly element-centric. In
part, this comes from the fact that management information models in this domain have not
been established and are in their infancy, in particular when it comes to manage OPT as a domain
in a network context, beyond the network elements. This paper therefore presents an OPT
information model. It addresses the need to model the network elements as well as the OPT
application domain as a whole, providing a level of abstraction that allows to hide much of the
underlying distribution of voice functionality. How the model can be leveraged by management
applications is also discussed.
OPT management is still a relatively new domain that has so far been largely dominated by
element-centric solutions. While adequate for smaller deployments, those solutions have
limitations dealing with the challenges resulting from the distribution of voice functionality and
openness, exposing much of that complexity to the user.

Network Operations and Management Symposium                                        Page 3
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

      A Service-Centric IP Quality of Service Architecture for Next
                          Generation Networks
            D. Goderis, S. Van den Bosch, Y.T’Joens Alcatel Network Strategy Group
    G. Pavlou, P. Trimintzios Centre Communication Systems Research, University of Surrey
                                    P. Georgatsos Algonet S.A.
               G. Memenios, E. Mykoniati National Technical University of Athens
          D. Griffin Dept of Electronic and Electrical Eng. University College London
                                C. Jacquenet France Telecom R&D
Today the Internet attempts to deliver traffic as soon as possible within the limits of its
abilities, but without any guarantees related to throughput, delay, inter-packet delay variation
and packet loss. So far this so-called best-effort forwarding paradigm has worked well because
most IP applications are low-priority and lowbandwidth data applications with high tolerance on
delay and delay-variation.
Value-added IP services, however, like Voice over IP (VoIP) and other multimedia applications,
require stringent end-to-end Quality of Service (QoS) guarantees.
Therefore, the key challenge for Next Generation Networks (NGNs) is to extend IP-based
networks with scalable, multi-service QoS capabilities, while still providing the key advantages
of IP that made the Internet possible. On these multiservice networks, operators will have to
honour complex Service Level Agreements (SLAs), acknowledging different types of traffic in
terms of bandwidth requirements, delay and other QoS parameters.
Within the Internet Engineering Task Force (IETF) several IP QoS technologies have been
proposed. Integrated Services (IntServ) was the first proposal, based on per-flow resource
reservation and admission control through the Resource reSerVation Protocol (RSVP). The main
disadvantage is that the required per-flow state information and QoS treatment in the core IP
network pose severe scalability problems. These problems led to the development of the
Differentiated Services architecture, which allows for flow aggregation in order to deal with
the scalability issues.
IP Differentiated Services is widely seen as the framework to provide Quality of Service (QoS)
in the Internet in a scalable fashion. However many issues have still not been fully addressed,
such as: the way Per-Hop Behaviours can be combined to provide end-to-end services; the
specification of admission control and resource reservation mechanisms; and the role of
management plane functionality and its integration with the control and data planes. This paper
presents the Service Management aspects of an integrated control and management
architecture for supporting end-to-end QoS-based IP services in Next Generation Networks.
The focus of the paper has been on the framework for the deployment of service and resource
management algorithms in an overall architecture, which meets the demands of future QoS-
based IP networks rather than on a detailed treatment of the specific algorithms and
mechanisms to be deployed. The main results are the decomposition of the overall problem into

Network Operations and Management Symposium                                       Page 4
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

specific functional blocks, the separation of service-related and resource-related sub-systems,
and the two level approach to service negotiation and invocation.

     A Proposal of QoS Control Architecture and Resource Assignment
                            K. Isoyama, H. Saito, Y. Nagao, M. Yoshida
                                  NEC Corporation, Hinode 1131
                                           Abiko, Chiba
                  k-iso@bc, h-saito@bx, y-nagao@bl, m-yoshida@em}

IP networks have evolved into a global and commercial communication infrastructure. The IP-
related technologies take a major part in technical bases and universal mechanisms not only for
internet traffic, but also for other communication services. These are telephone service which is
sensitive to delay, and financial communication service which requires reliability. Also bandwidth
trading service which needs guaranteed bandwidth is emerging. Users/Applications require
different Quality of Service (QoS) and some may require a QoS guarantee.
As internet service becomes commercial, providers need to charge for their services, and to
prescribe services to provide. A Service Level Agreement (SLA) is a contract between a user
and a provider, which defines service specifications and guarantees/responsibilities.
Consequently, providers need to control their networks based on SLAs.
To support a variety of applications with different characteristics and requirements at adequate
service levels, QoS technologies and their enabling mechanisms are expected to be developed
and deployed; Multiprotocol Label Switching (MPLS), Differentiated Service (Diffserv) and
traffic engineering are typical candidates.
MPLS architecture introduces the concept of connections into IP networks. The MPLS system
forwards packets based on the label assigned in each packet. In MPLS architecture, providers
can control Label Switched Paths (LSPs) explicitly, so that they can achieve traffic engineering
with load distribution to paths within the network.
Diffserv architecture controls QoS for each class of aggregated flows. Edge routers in a
Diffserv domain classify incoming packets, and assign each packet a class identifier, Diffserv
Code Point (DSCP). Core routers within the domain read DSCP values and forward the packets
based on Per Hop Behavior (PHB) of forwarding scheduling definition given for each class.
Consequently, the Diffserv architecture, which is class-based control instead of flow-based, can
provide scalable QoS even in a large-scale network.
Although these technologies take local QoS management into account, adequate mechanisms are
to be developed to control network-wide administration, provisioning, monitoring and calculation

Network Operations and Management Symposium                                        Page 5
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

of optimal QoS parameters to satisfy end-to-end QoS demands. Traffic engineering is a
technology that optimises network resources usages, load balances and performances.
Network architectures that integrate voice, video and conventional data into IP networks have
been proposed. However, some provide integrated service without any QoS, and others provide
only differentiated or prioritized QoS using mechanisms such as Diffserv. They are not
sufficient for guaranteed service.
For guaranteed QoS based on SLA, mechanisms that calculate adequate resource assignment and
parameters of QoS devices in the network are needed. In addition, mechanisms that manage and
monitor SLAs are also needed.
A policy control framework [6] is a mechanism, which provides a coherent resource assignment in
the whole network by applying centralized management. The policy framework allows network
operators to set up policy rules. However, since the policy framework controls resource
assignment based on fixed policy rules which are defined by network operators, it is difficult to
provide policies which are adaptive to change of network state and of amount and requirement of
user traffic.
This paper proposes a QoS Control architecture, which provides optimal resource assignments
(i.e., paths, bandwidth and buffers) to QoS applications in IP networks where various
applications are going to be introduced.
The objective of traffic engineering is to guarantee QoS of user/application traffic while
improving accommodation efficiency by optimising network resource usages. In the proposed
architecture, a QoS server calculates and provides optimal resource assignments on the basis of
information on the user/application resource request, state of network routing, resource
availability, and utilisation.
As future work, distributed or hierarchical server architecture should be studied in order to
acquire more scalability.

                Value Added Service Management in 3G networks
                              N. Houssos, E. Gazis and S.Panagiotakis,
                                       University of Athens
                                      151 26 Athens - Greece
                                    S. Gessler and A. Schuelke,
                                NEC Europe Network Laboratories
                                         Adenauerplatz 6
                                   69115 Heidelberg - Germany
                                          Sandra Quesnel
                                      Thales Communications
                                       66, rue du fosse blanc
                                   92 231 Gennevilliers - France

Network Operations and Management Symposium                                       Page 6
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

The forthcoming era of 3rd generation (3G) mobile communications is heralded to change the
wireless telecommunication experience dramatically. The development of 3G systems will be a
major step towards the convergence of the telecommunications and the Internet industries and
will result in the creation of an open and highly competitive network services marketplace, where
3rd party service developers, often termed Value-Added Service Providers (VASPs), will be able
to market their services to users of wireless and wired networks in a globally unified way1. Thus,
the efficient delivery of applications to end users will be more complex than ever before,
bearing also in mind the growing demand for ubiquitous, personalised and terminal aware service
provision. In this context, the task of service management becomes significantly more
demanding, as well as critical for the continuous growth and market success of the converging
worlds of telecommunications and the Internet. This paper proposes a potential solution to this
issue, by introducing a distributed software framework for the efficient management and
intelligent provision of value added services in 3G networks.
In mobile telecommunication environments until 2G, network access and service/application
provision has been statically bundled together in various subscription-based offerings by
network operators. The introduction of 3G systems, such as UMTS, which distinguish between
the network and the service/application domain, renders such strategies inefficient. To “make
3G happen” a critical mass of applications and services that are novel, attractive and exploit the
unique capabilities of 3G networks to the benefit of the user’s experience is necessary. In
contrast to the prevailing 2G norm, where services are developed by a limited number of
telecommunications experts (e.g., operators, equipment vendors), 3G applications should be
substantially contributed by a large variety of non-telecomm literate market players, including
software vendors and content providers. This fact is expected to foster competition and
differentiated service offerings and to accelerate system deployment and service take-up, thus
minimizing investment risks. It is therefore envisaged that the evolution of mobile
telecommunications towards 3G – and beyond it – will be driven by two major forces:
 - A multitude of services with rich functionality that will provide the revenue stream required
 to amortize the high 3G investments.
 - The existence of multiple business players in the value chain that will cooperate in the
 provision of enriched service offerings.
In the context described before, there is a shift in the traditional paradigms of management
from managing network elements and networks to managing services. The abundance of services
that will be typically developed by many co-operating entities and will need to be rapidly
deployed on multiple types of networks significantly complicates service deployment and
provisioning compared to 2G networks. Moreover, the requirement for QoS provisioning and for
sophisticated, flexible billing schemes makes the service management task more difficult than it
is in the Internet world. Therefore, a need is emerging for intelligent and flexible service
provision platforms that will mediate between service developers, network operators and end-
users and provide a set of functionality not available until now.
To be made globally available, 3G services – typically – developed by independent application
providers, must adhere to a deployment model that is network and terminal independent.

Network Operations and Management Symposium                                        Page 7
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

Straightforward and automated deployment procedures will be a critical feature for the
successful lifecycle management of 3G services. Its significance lies in the fact that it enables
marketing a large number of services whilst reducing service time-to-market, two of the most
important business requirements of 3rd generation networks. The service deployment process
unfolds itself into a number of distinct procedures:
This paper investigated issues related to service management in the new world of 3G mobile
communications and the emerging requirements that make this task significantly more
complicated that in the past. An advanced service management and provision framework was
introduced as a solution to these issues, aiming also to be complementary and value-adding to
relevant standardization activities. In particular, the architecture and functionality of the
central platform component was presented in more detail and the implementation of a prototype
of the software platform, developed in the context of the MOBIVAS project was briefly
Work presented in this paper has been performed in the framework of the project IST
MOBIVAS, which is partly funded by the European Community.

     Shifts In Network Management Paradigms Arising From The New
                        Competition Environment
                                           Decio Ongaro
                                     Tecnosistemi Group S.p.A.
                                       Strada 4, Palazzo Q3
                                    20089 Rozzano Milano, Italy
                                   decio.ongaro@tecnosistemi .it
The increasing competition in offering telecom services has driven the multiplication of the
number of network and service operators which have entered the market.
It has also driven the technologies and services deployed in an effort to differentiate offers.
The passage from voice-only offerings to data/multimedia has as a consequence also the
multiplication of the "core skills" necessary to be competitive in these new areas. This also is a
factor in segmenting the operators/service providers, leading to a further multiplication of the
players involved.
The operator/customer relationship, clearly the fundamental issue in such a competitive
environment, has to be conjugated with the actual complexity of this world.
All networks are globally interconnected. Each customer expects connectivity and services not
just from its own operator of choice, but throughout and across all networks and services.
Thus a large extent of the entities actually contributing to the services actually used by a given
customer are not under the direct control and supervision by its operator of choice, as was an
established fact in the “traditional” telecom world which has all but disappeared.

Network Operations and Management Symposium                                        Page 8
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

Mobility and the extensive penetration achieved by cellular networks like GSM where subscriber
roaming is a key factor complicates this picture of satisfying subscriber needs even more.
Nowadays competition involves local operators, national operators, international operators.
The current phase of the “industry cycle” indicates that many such operators will not survive, at
least as independent entities. Many of today’s independent entities will either disappear or be
absorbed by the winners.
However, even at the end of a consolidation process a manifold competition of network operators
will continue to exist in each market and location.
All these networks will be thoroughly interconnected, as the telecom market is by nature global.
“Global” are the customer’s expectations.
The scenario becomes even more composite when one includes “collective” customers, such as
industries and organizations. Each of these runs its own private IT and voice network, often in
different locations. These are connected and interconnected with “public” operators’ networks
and with service providers –often several of them.
These collective customers naturally tend to use a more extensive and sophisticated variety of
services than individual (consumer) subscribers. Quality requirements are also more
sophisticated, which makes the impact on maintenance and operation issues are consistently
Even if one restricts to voice-only, the passage to wireless cellular services has marked a
substantial increase in complexity, due to the impact of issues related to mobility.
One such issue is that of roaming across networks, already noted. Another is the fact that the
patterns of traffic peaks have become a highly geo- and time-dynamic matter, and also
dependent on the launch of new tariffs. As a consequence, dynamic detection of network
stresses is key to network optimisation.
Even if voice services are still the core revenue generator, they are not a market differentiator.
Focus on data and multimedia becomes more and more apparent….
The purpose of the present paper is to review and comment the impact of this scenario of
changing environment in the scope of Network Operation and Maintenance, with comments on
the shifts in relative importance of solutions to be implemented to cater to these new

        Distributed Networking System for Internet Access Service
                      Daniel W. K. Hong, J. G. Song, J. H. Yoo, and W. S. Kim
                            ANL, R&D Group, KT Daejeon 305-811 Korea
                                           C. S. Hong
                         Kyung Hee University -Kyungki-Do 449-701 Korea
Network Operations and Management Symposium                                        Page 9
Florence, Italy, 15-19 April 2002
                                         NOMS 2002

As a widely accepted telecommunications system, the forthcoming B-ISDN and its packet-
oriented transmission technology in ATM have many promising new telecommunication services,
such as the multimedia and high-speed data transfer.
Those services demand higher bandwidth connectivity that can be transferred through fiber
optic or coaxial cables. Even though Plain Old Telephone System (POTS) copper cables offer a
lower cost both in telecommunications and data communications, the existing modems can’t
accommodate the forthcoming high-speed multimedia services.
Therefore, a new platform that is capable of accommodating such speed is necessary.
The Asymmetric Digital Subscriber Loop (ADSL) technology is a new platform for delivering
broadband services to homes and small businesses. ADSL can be implemented on most of the
existing copper infrastructure, enabling the rapid and near ubiquitous offering of new high-
speed Internet access services at minimal cost.
The number of ADSL subscribers in Korea Telecom is remarkably increasing by as many as eight
hundred new subscribers every day, with a total number now already at 3 million. Because of
such a large number of customers, it is important to keep these customers satisfied. Customer
satisfaction is the key to success in today’s competitive environments. In addition, the success
of a network provider depends entirely on its capability of bringing collocation facilities on line
easily, transmitting data flawlessly, and pinpointing problems instantly to avoid service
To fulfil these requirements, the authors propose a systematic networking system that can
easily accommodate the explosive growth in the number of ADSL subscribers and can efficiently
and uniformly manage the various heterogeneous network equipments that are increasing in
proportion to the growth of ADSL subscribers. This paper defines the scaleable network model
that can easily accommodate the excessive growth of subscribers. In addition, they define the
distributed networking system architecture aiming to maximize the manageability of the various
heterogeneous network elements that are needed to provide Internet access via ATM over
With the proposed distributed networking system, Korea Telecom can accommodate over fifteen
thousand ADSL subscribers per day without any interruption or mediation from the operator,
and can systematically manage the Internet access service using the ATM over ADSL
In addition, the various heterogeneous network elements provided by different vendors using
the unified element adaptation interface can be uniformly managed. The new management
functions can easily be added to the existing distributed networking system and the existing
network management functions can easily be revised. Now, this network system has been tested
under the real network environment. After finishing this field trial, the valuation results, open
issues and technical challenges of this model will be published.

Network Operations and Management Symposium                                        Page 10
Florence, Italy, 15-19 April 2002

To top