Catálogo de publicaciones - libros

We propose hierarchically distributed path computation elements (HDPCEs)that manage a multilayered GMPLS network and provide traffic engineering paths for end-to-end routes. We apply HDPCEs to provide a VPN service on optical networks for IP customers. One of the VPN services that HDPCEs provide is individual-flow-based bandwidth allocation, which assures end-to-end QoS. However, conducting flow-based bandwidth assurance has been difficult because of the heavy burden on the admission controller (AC) and its routing delay. We made conducting flow-based bandwidth assurance possible by distributing ACs and advanced route creation. Another service provides resilient paths that are free from failure in the optical domain. There is no clear SLA on an individual path basis for the path backup policy. HDPCEs provide backup methods, such as protection, preplanned restoration, and active restoration according to the SLA with the customer.

For QoS guaranteed DiffServ-over-MPLS connection provisioning in multi-domain networks, the managed objects in each domain should be globally standardized and publicly accessible by other NMSs. And each NMS participating in the inter-domain networking should negotiate with other NMSs for end-to-end service provisioning. In this paper, we propose SLA negotiation by COPS protocol and design the managed objects by expanding existing experimental DMTF CIM MOFs. We propose service location protocol based directory agent to register/discover the NMSs’ that participate in the multi-domain TE-LSP connection provisioning. For better performance, we design the multi-threaded provider implementations on OpenPegasus WBEM platform. The detailed implementations of the providers with performance analysis are explained.

Recently, new versions of the TCP protocols such as HSTCP (High Speed TCP) has proposed to adopt the rapid growth of the Internet. However, it is important to consider a deployment scenario to shift gradually from the current to the new versions of TCP. A fairness problem between the performance of both current and new versions of TCP will be arisen when they share the same bottleneck link, which leads some complaints from people using the current version of TCP. In this paper, we first clarify the relation between the deployment share of HSTCP and the negative impact to TCP Reno users. We then propose a new management mechanism which supports TCP version migration from TCP Reno to HSTCP based on the relation. Through simulation, we also show that our mechanism enables to accelerate a gradual migration of TCP versions by achieving the end users’ complaints as little as possible.

This paper intends to introduce the development of a terminal agent for QoS measurement that is suitable for an NGN environment, and to summarize the results of its performance test. Using the terminal agent for service quality measurement, it is possible to measure the QoE-based end-to-end service quality of voice phone, video phone, and VoD services. In addition, the terminal agent, installed in the user terminal (IP-based audio and video phones), as a software or hardware chip, measures the quality index for voice and video related multimedia services, such as R-value, MOS value, call success rate, one-way delay, jitter, packet loss, resolution. The terminal agent also applies the packet capturing method when using the actual service, and analyzes SIP, RTP, and RTCP protocol headers of the received packet. When the terminal agent was implemented in the IP video phone and compared with the performance of a general QoS measurement tool, there were barely any errors in the measurement value. It has also been identified that end-to-end service quality management, which is close to QoE, was possible.

Peer-to-Peer systems are based on the concept of resources localization and mutualization in dynamic context. In specific environment such as mobile networks, characterized by high variability and dynamicity of network conditions and performances, where nodes can join and leave the network dynamically, resources reliability and availability constitute a critical issue. To deal with this critical issue, we introduce a new concept and model called ”P4L” () which define a novel P2P architecture, aims to improve: fault-tolerance, self- organization and scalability, with limited complexity while providing a rapid convergence for the lookup algorithm. The cost of ”P4L” lookup is (∑ln ()) where is the number of nodes on ring level (). ”P4L” is efficiently adapted to the context where nodes join and leave dynamically and frequently. Each node maintains routing information of 2*(ln ()), where is the number of nodes on one ring. Thus ”P4L” is well adapted for terminals with limited resources such as mobile terminals. ”P4L” is based on ring topology with each ring connecting ”neighbouring” nodes in terms of physical and logical position. When ”P4L” is combined with broadcast mechanism, the lookup process is significantly improved. The proposed model is evaluated and compared with Chord protocol, an extension is proposed to support IPv6.

In this paper, we propose a new approach to secure and easy-to-use remote access to networked appliances (NAs). Based on the proposed approach, we develop a system in which servers in an Internet service provider network and a residential gateway (RGW) in a home network play an important role, including secure communication. The system allows us to access NAs, using mobile phone and PC anytime and anywhere. Our emphasis is that the system is based on zeroconf protocols including UPnP, and only by connecting an NA to a RGW enables secure remote access. Another emphasis is deployability. A RGW that has been widely deployed in home networks and always operates for VoIP call serves the zeroconf protocols. Interoperability with both UPnP compliant and noncompliant NAs, together available in the marketplace, is also considered. We deployed the system and conducted a field trial in which approximately 1700 users took part in for four months. Empirical results in terms of the zeroconf protocols will be shown with future directions.

As a network & service provider, KT has long felt the necessity of integrating many of its OSS (Operation Support System) systems, and in 2003, KT finally developed a new system to serve its various maintenance activities. This system is called the NeOSS (New Operations Support System) System, which uses XML (Extensible Markup Language), EAI (Enterprise Application Integration) and Web service based on Microsoft’s .NET platform technology. This paper shows the system architecture of the NeOSS-ADM(Access Domain Management), which is going to be developed using the .NET environment, service maintenance framework and the method of maintenance for NEs (network elements) for base BS. And this system will give more high quality service to customers and reduce the customer’s complaints.

Systems designed to be self-healing are able to heal themselves at runtime in response to changing environmental or operational circumstances. Thus, the goal is to avoid catastrophic failure through prompt execution of remedial actions. This paper proposes a self-healing mechanism that monitors, diagnoses and heals its own internal problems using self-awareness as contextual information. The self-management system that encapsulates the self-healing mechanism related to reliability improvement addresses: , in order to perform self-diagnosis and self-healing. To confirm the effectiveness of self-healing mechanism, practical experiments are conducted with a prototype system.

As the network technology has been developed, the Next Generation Internet (NGI) such as Internet2, KOREN, KREONET2 and etc has been deployed to support bandwidth of Giga bps. And, various applications such as the video conference, the tele-surgery and etc that require high bandwidth has been developed and operating on the NGI, especially KOREN and KREONET2 in Korea. When the applications are operating and happen to face some problems, various tools for traffic measurement are usually used to analyze them, and provide traffic information in each link on total flow path or other metrics such as end-to-end delay, jitter and etc in most cases. However, although with these tools most of network problems can be properly analyzed, the problem in the user’s view point can not be resolved so that which part of networks and/or which ones among user’s systems and nodes on the flow path cause the problems for the user’s flow is not discriminated Therefore, this paper designs a end-to-end performance monitoring system for the user’s flow that the user can access performance data for user’s flow on the flow route, and describes the experience about deployment on KOREN.

In convergence telecommunication environment, Business Agility plays very important role in the OSS(Operation Support System) when telco provide new merged services to customer on time. But, the OSS also becomes more and more complicated to know even what part of it should be fixed for adopting new services. This paper proposes SOA-based OSS architecture for telecommunication services in order to cope with this situation. We present the designing method of services of SOA and architecture for OSS by investigating the architectural issues of the unit of derived service elements from OSS and designing the most suitable architecture of it. By adopting the represented architecture for OSS, telco can provide new convergence service to customers faster than the competitor on the market.