A Broadband Network Management Practice: KT Internet 2006.9.28 Jinhyoun Youn (jyoun@kt.co.kr) KT Network Strategic Group/ Internet Department

Contents Brief History of KT Internet Service and NMS KT Internet Management Systems Experiences NeOSS-IP Core NMS Performance Measurements Traffic Analysis Routing Information Management New Features for Premium Networks Future Directions

Brief History of KT Internet Service

Brief History of KT Internet NMSs At the initial stage(~1996), primitive monitoring and fault management functions based on the production systems like Openview or CiscoWorks were the main usage of Internet NMS As the network grew, in-house software were much more preferred and KOSMOS(Kornet Service Managements and Operation Support systems,1996) was the first integrated style Internet NMS for KT. Simple monitoring functions were expanded to reacheability and availability management As the portion of the Internet (and data networks) grew in the revenue, the interoperability with other infrastructure (espeially the main telco infrastructure) got much more important, and NEOSS-IP was confronted in 2002. Various Measurements and analysis were added.

KT Internet Service and Mgmt NeOSS-IP and its sub systems CORE-NMS, Premium NMS, Multicast NMS, Access NMS Analysis Systems

NeOSS-IP and its Subsystems NeOSS-IP is an NMS system for IP networks in KT NeOSS systems NeOSS-IP includes subsystems for operation tools like Traffic, Performance, RIB analysis systems Configurations, Remote Operations support tools

Performance Measurements IPAS(Internet Performance Analysis System) As a start point of service level (quality) managements Like sorts of network weather systems UDP/TCP active measurements over thousands of paths per minutes using probes (special-purposed routers) Focus on UDP and jitter delay and loss rather than simple ICMP stuffs

Performance Measurements(2) What we can learn from IPAS General Packet Transmission or IP Service Quality Levels over the backbone networks Abrupt packet transmission delays due to events on the backbone networks

Traffic Analysis Three ways of Traffic Understanding General traffic volume analysis using CORE-NMS Traffic flow analysis using routers and IP-FLOW system Traffic Packet Analysis using PACAS(Packet Characteristics Analysis system)

Traffic Analysis (2) What we can learn from CORE-NMS The total traffic volumes over KT backbone, overseas, IX networks The link utilizations What we can learn from IP-FLOW The general trend of the traffic usage ex. per applications(ports), per ASN, per POPs/Access services and so on

Traffic Analysis (3) What we can learn from PACAS The analysis results can be used directly for engineering Illegal packets with wrong IPs DDOS packets Out-of-Sequence/Order ratio and so on Packet Level Characteristics PDF of Packet Arrivals Traffic volumes at tiny time units(micorseconds units) Microscopic version of data which we can get from IP-Flows i.e. per ports/IP Addrs/ASN traffic distribution and so on

Routing Information Management IRIMS(IP Routing Information Managements System) Routes(Prefixes) changes archives Realtime change monitoring supported Routing Daemons and SNMP are used together

Routing Information Management(2) What we can learn from IRIMS Real time routing information change monitoring BGP Session and prefix changes monitoring ISIS links changes monitoring FLAPS and routing process overloads Inconsistent AS

New Features for Premium Networks MPLS managements New elements for MPLS: LDP, LSP, LIB, MPLS IFs and Tunnels Dynamic LSP managements MPLS-TE Prototypes QoS Managements New elements for QoS(DIFF_SERV) : Queues and policies DIFF_SERV management with backbone policy server Per class Traffic, Delay, Jitter, Loss Per Tunnel Traffic Multicast Managements New elements for Multicasts: Multicast tunnels, RPs, Ifs, IGMP, PIM_SM, MSDP Management

Near Future: Beyond Broadband Service Needs More bandwidths Various access technology Personalization Convergences in services BCN Service Concepts Broadband Convergence Network high-quality broadband multimedia services Integrating telecom, broadcasting, and Internet seamlessly at anywhere, anytime, and using any devices KEYWORDS: QoS, IPv6, MPLS, IMS, Convergence

BCN at the moment BcN(backbone network-the premium network) has been deploying since 2005 DIFFSERV for the basic backbone CoS MPLS for the enhanced availability Packetized Voice, IP-TV, and Wibro on the premium network CoS applied for service classes BcN Service Infrastructure including various control mechanisms will be implemented soon IMS based service control systems: Resources and Services(Session) Large-scale user(subscribers) handling features

NMS Expectations AS IS BROADBAND and its components have been the main NMS objects so far Mainly networks, elements, and facilities are main concerns for NMS To Be Ubiquitous service and convergence might be regarded as the next theme End-to-end application/service supports will be the next concerns The volumes, numbers, core-technologies of the networks will be continuously increasing The underlying network functions and the services will be tightly coupled and the layering conventions will be ambiguous

Concluding Remarks : What we will do Even though the underlying technologies or networks is expected to be more complex, the NMS and its view for operators should be simpler: Integrated but selectable(and flexible) user interfaces On the shared platforms, the interoperability between NMS systems for various networks/services should be achieved. To support convergence trends, the unified platform for supporting customers will be more important.