cs158b-class-04-nms.pptx

CS158B: Network Management Fall 2024 Network Architectures Network Management Framework Copyright Arun Saha, 2024. All rights reserved. 1 Kinds of Networks Copyright Arun Saha, 2024. All rights reserved. 2 Campus Network Copyright Arun Saha, 2...

CS158B: Network Management Fall 2024 Network Architectures Network Management Framework Copyright Arun Saha, 2024. All rights reserved. 1 Kinds of Networks Copyright Arun Saha, 2024. All rights reserved. 2 Campus Network Copyright Arun Saha, 2024. All rights reserved. 3 Copyright Arun Saha, 2024. All rights reserved. 4 Copyright Arun Saha, 2024. All rights reserved. 5 Data Center Network Copyright Arun Saha, 2024. All rights reserved. 6 Evolution of Data Center Network Copyright Arun Saha, 2024. All rights reserved. 7 Service Provider Network Copyright Arun Saha, 2024. All rights reserved. 8 Copyright Arun Saha, 2024. All rights reserved. 9 Cloud Network Copyright Arun Saha, 2024. All rights reserved. 10 FCAPS model Copyright Arun Saha, 2024. All rights reserved. 11 FCAPS model Fault Configuration Accounting Performance Security Copyright Arun Saha, 2024. All rights reserved. 12 Fault Management Goal: Detection, Isolation, and Correction of abnormal operation Fault := An abnormal condition that requires management attention (or action) to repair Fault != Single Error Example: Link down, Port down, Device down Copyright Arun Saha, 2024. All rights reserved. 13 Configuration Management Goal: Configure all hosts and devices for their intended behavior Drives network functionality Originally, almost always human involved Nowadays, could be either Human driven, or Controller* Driven Controller: Higher Level Management Software Example: Port disable/enable Configure VLAN Configure tunnel Copyright Arun Saha, 2024. All rights reserved. 14 Accounting Management Goal: Gather all kinds of data from all hosts and devices Alternate names: Stats / Metrics / Telemetry Example: Bytes sent/received, Packets sent/received Copyright Arun Saha, 2024. All rights reserved. 15 Performance Management Goal: Ensure network performance remains at acceptable levels Tracks Network Health, Quality-of-Service, User Experience Example: Throughput, Latency, Loss rate, Utilization SLA/SLO Copyright Arun Saha, 2024. All rights reserved. 16 Security Management Goal: Keep network secure All access to the network are authenticated and authorized Generation, Distribution, Storing Encryption keys Example: Security Rules Copyright Arun Saha, 2024. All rights reserved. 17 Network Management System Copyright Arun Saha, 2024. All rights reserved. 18 Network Management System (NMS) Collection of tools for network configuration and monitoring that is integrated and unified in the following senses: A single operator interface with a powerful but user- friendly set of commands for performing most or all network management tasks A minimum amount of separate equipment is necessary That is, most of the hardware and software required for network management is incorporated into the existing user equipment Copyright Arun Saha, 2024. All rights reserved. 19 Network Management Entity (NME) Each device contains a collection of software devoted to the NM task, often known as NME. Each NME performs the following: Collect stats on communications and network-related activities Store stats locally Respond to commands from the Network Control Center (NCC), e.g., Transmit collected stats to the NCC Change a parameter (e.g., a timer) Provide status info (e.g., parameter values) Generate artificial traffic to perform a test Send message to NCC when local conditions undergo significant change Copyright Arun Saha, 2024. All rights reserved. 21 Network Monitoring Information STATIC: Information that seldom changes, e.g., router id, number of ports DYNAMIC: Related to the events in a network, e.g., forwarding state of a port in a switch STATISTICAL: Frequently changing information, e.g., number of bytes sent/received, number of packets sent/received Copyright Arun Saha, 2024. All rights reserved. 22 Network Monitoring Configuration Monitoring Application: Functions of network monitoring that are visible to the user, e.g., faults, performance Monitoring Application Manager function: Retrieves information Manager from other elements Function Agent function: Gathers and records management information for one or more network elements and communicates that to monitor Agent Function Managed objects: Management information representing resources and Managed Objects their activities Copyright Arun Saha, 2024. All rights reserved. 23 … and optionally Monitoring Application Manager Monitoring agent: Function Aggregates and possibly summarizes Agent to the Monitoring Monitoring Agent Application Manager to the Agent Function Agent Function Agent Function Managed Managed Objects Objects Copyright Arun Saha, 2024. All rights reserved. 24 Polling Polling is a request-response interaction between the manager and the agent Copyright Arun Saha, 2024. All rights reserved. 25 Event Reporting With event-reporting, the initiative is with the agent The manager is in the role of listener, waiting for incoming information A report could be: periodic with a preconfigured interval … that can be modified by the manager significant event (e.g., state change) unusual event (e.g., a fault) Copyright Arun Saha, 2024. All rights reserved. 26 Performance Monitoring Indicators Service-Oriented Availability: % of time network or application is available to the users Response time Accuracy Efficiency-Oriented Throughput: events per unit time Utilization: % of theoretical capacity of a resource Copyright Arun Saha, 2024. All rights reserved. 27 Performance Quest Is traffic evenly distributed among the network hosts or are there source-destination pairs with unusually heavy traffic? What is the percentage of each type of packet? Are some packet types of unusually high frequency, indicating an error, or an inefficient protocol? What is the distribution of data packet sizes? Which host, port is the biggest source of traffic? Which host, port is the biggest sink of traffic? Copyright Arun Saha, 2024. All rights reserved. 28 Fault Monitoring Problems: Unobservable faults, e.g., process internal Copyright Arun Saha, 2024. All rights reserved. 29 Problem What is the maximum yearly downtime for 3-nine SLA? I.e., SLA = 99.9% uptime Copyright Arun Saha, 2024. All rights reserved. 30 Solution Copyright Arun Saha, 2024. All rights reserved. 31 Problem There are 100 computers in a cluster. The probability of failure of a single computer is 0.1%. Assuming all computers are independent, what is the probability that there is at least one failure in the cluster? Copyright Arun Saha, 2024. All rights reserved. 32 Solution Let p = Probability of a Single Computer Failure = 0.1% = 0.001 Let n = Total number of computers in the cluster = 100 P(no failures) = (1-p)n P(at least one failure) = 1 – P(no failures) = 1 – (1 – 0.001)100 = 1.- 0.905 For n = 1000, = 0.095 P(at least one failure) = 9.5% = 63% Copyright Arun Saha, 2024. All rights reserved. 33