Once a network grows beyond a few devices, someone has to know whether it is healthy, who is using it, what changed and what failed. Network management is the set of tools and practices for monitoring, configuring and troubleshooting networks. The classic protocol, SNMP, is still in use, and its ideas live on in Prometheus, Grafana, NetFlow, streaming telemetry and AIOps. This article revisits the five units of the JNTUK course and connects each to modern monitoring.
Unit 1: Network management fundamentals and architecture
Why manage networks
Managed networks are more reliable, secure and efficient, and problems are found before users report them. A network management system has to cope with heterogeneous devices from many vendors, so it relies on standard protocols.
The FCAPS model (ISO)
| Area | Question | Examples |
|---|---|---|
| Fault management | What is broken? | Detect, isolate and fix faults; alarms, trouble tickets, root-cause analysis |
| Configuration management | What is configured? | Inventory, device configuration, version control, backups, provisioning |
| Accounting (administration) | Who uses what? | Usage metering, billing, quotas |
| Performance management | How well does it run? | Throughput, latency, errors, utilisation, capacity planning |
| Security management | Is it protected? | Access control, authentication, audit logs, intrusion detection |
Management architecture
A manager–agent model:
| Component | Role |
|---|---|
| Network management station (NMS) / manager | Software that monitors and controls; the operator’s console |
| Managed device | Router, switch, server, printer, UPS |
| Agent | Software in each device that collects data and answers the manager |
| Management information base (MIB) | The database of managed objects in each agent |
| Management protocol | The language between manager and agent (SNMP, CMIP, NETCONF) |
Models: centralised (one NMS), distributed (several NMSs, each for a domain), and hierarchical (managers of managers). Management can be in-band (over the same network) or out-of-band (a separate network, which is useful when the main network is down).
Standards and models
| Model | Origin | Note |
|---|---|---|
| SNMP (Internet) | IETF | Simple, widely used |
| OSI / CMIP | ISO | Rich but complex; little used outside telecom |
| TMN | ITU-T | Telecommunications management |
| TMF Frameworx, ITIL | Business and service processes |
Today: FCAPS is still the best checklist for a monitoring programme: alerts (fault), IaC and config drift detection (configuration), cost allocation (accounting), SLO dashboards (performance) and audit logs (security). The manager-agent idea is the same as Prometheus scraping exporters, or a monitoring agent on each server.
Unit 2: SNMP and the MIB
The Simple Network Management Protocol
SNMP is the standard for TCP/IP network management. It runs over UDP (port 161 for agents, 162 for traps) and is deliberately simple.
| Operation | Purpose |
|---|---|
| GetRequest | Read the value of a specific object |
| GetNextRequest | Read the next object in the MIB order (used to walk a table) |
| SetRequest | Change a value (configure a device) |
| GetResponse / Response | The agent’s reply |
| Trap | An unsolicited alert from the agent (link down, cold start) |
Two ways to collect data: polling (the manager asks at intervals) and traps/notifications (the agent tells the manager). Polling gives regular trend data; traps give fast alerts. Communities (SNMPv1 and v2c) are text “passwords” such as public (read) and private (write), sent in clear text, a serious weakness.
Structure of Management Information (SMI)
SMI defines how managed objects are described: with ASN.1 (Abstract Syntax Notation) and encoded with BER (Basic Encoding Rules). Object types include INTEGER, OCTET STRING, OBJECT IDENTIFIER, Counter, Gauge, TimeTicks and IpAddress. A Counter only increases and wraps; a Gauge can go up and down.
The MIB
The Management Information Base is a tree of objects, each identified by an Object Identifier (OID), a dotted sequence of numbers:
1.3.6.1.2.1.1.1.0 = iso.org.dod.internet.mgmt.mib-2.system.sysDescr.0
| MIB group (under mib-2) | Content |
|---|---|
| system | Device description, uptime, contact, name |
| interfaces | Per-interface status, speed, byte and error counters |
| ip, icmp, tcp, udp | Protocol statistics |
| snmp | The agent’s own statistics |
MIB-II is the standard set; vendors add their own under the enterprises branch (1.3.6.1.4.1). Tables are indexed (for example, one row per interface). Tools such as snmpget, snmpwalk, snmpset and MIB browsers explore them.
Today:
snmpwalk -v2c -c public 192.0.2.1is still used to read routers, switches, UPS units and storage arrays. Prometheus’s snmp_exporter, Zabbix, LibreNMS, Nagios and PRTG poll SNMP and turn counters into graphs. Counters are turned into rates (bytes per second), just likerate()in PromQL.
Unit 3: SNMPv2, SNMPv3 and RMON
SNMPv2 (v2c)
Improvements over SNMPv1: GetBulkRequest (fetch many values efficiently), InformRequest (acknowledged notifications, also manager to manager), 64-bit counters (needed for fast links, where 32-bit counters wrap in seconds), and better error handling. It still uses community strings, so security was not fixed.
SNMPv3
SNMPv3 adds the security and administration that earlier versions lacked.
| Feature | Meaning |
|---|---|
| User-based Security Model (USM) | Named users with authentication (HMAC-SHA) and optional privacy (encryption, AES) |
| Security levels | noAuthNoPriv, authNoPriv, authPriv |
| View-based Access Control Model (VACM) | Controls which user may read or write which part of the MIB |
| Message integrity, replay protection | Timeliness checks |
| Modular architecture | Engines, applications and a pluggable security model |
Use SNMPv3 with authPriv wherever possible; if v2c is unavoidable, restrict it by network and use read-only communities.
| Version | Security | Features |
|---|---|---|
| v1 | Community string | Basic |
| v2c | Community string | Bulk, Inform, 64-bit counters |
| v3 | Authentication and encryption, access control | Same operations as v2 plus security |
RMON (Remote Monitoring)
SNMP polling gives device-level counters. RMON puts monitoring probes on network segments that collect statistics and traffic analysis locally, so the manager does not have to poll constantly.
| RMON1 group | Content |
|---|---|
| statistics | Packets, errors, size distribution per interface |
| history | Samples over time |
| alarm, event | Thresholds that trigger alerts or logs |
| hosts, hostTopN, matrix | Per-host and conversation statistics, top talkers |
| filter, capture | Packet filtering and capture |
RMON2 extends monitoring to the network and application layers (protocol distribution, address mapping). Related traffic technologies: NetFlow, sFlow and IPFIX (flow records: who talked to whom, how much, which port).
Today: flow data (NetFlow, VPC Flow Logs in AWS, GCP flow logs) answers “which service or IP is using the bandwidth?”. Packet capture (
tcpdump, Wireshark) is the descendant of RMON capture. SNMPv3 is required in many security baselines, and a firewall or switch that still only offers v1 or v2c should be isolated on a management network.
Unit 4: Telecommunications and broadband network management
TMN (Telecommunications Management Network)
TMN from the ITU-T manages telecom networks in layers:
| Layer | Concern |
|---|---|
| Business management | Business goals, policies |
| Service management | Service quality, customer interface, billing |
| Network management | Whole-network view, end-to-end paths |
| Element management | Individual equipment, with element managers |
| Network element | The physical devices |
It uses the OSI CMIP and object-oriented models, with Q3 interfaces between layers. It is the basis for how operators run large multi-vendor networks.
Broadband and access networks
Network management extended to the technologies of carrier networks:
| Technology | Management concerns |
|---|---|
| ATM and SONET/SDH | Virtual circuits, performance monitoring, protection switching |
| DSL, cable and fibre access (PON, GPON) | Line quality, subscriber provisioning |
| MPLS and carrier Ethernet | Path setup, QoS, service level agreements |
| Mobile networks (4G and 5G) | Radio resource management, handovers, subscriber data |
Functions: service provisioning, SLA monitoring, trouble-ticket systems and OSS/BSS (operations and business support systems) integration.
Modern network management protocols
| Protocol | Idea |
|---|---|
| NETCONF | XML-based configuration over SSH, with transactions and rollback |
| RESTCONF | The same model over HTTPS and REST |
| YANG | A data modelling language that describes configuration and state, the successor to MIBs |
| gNMI / gRPC streaming telemetry | Devices push structured data continuously instead of being polled |
| Syslog | Event messages from devices to a log server |
Today: operators and large enterprises automate with NETCONF and YANG models, and use streaming telemetry for second-level visibility. Cloud networks are managed through APIs and infrastructure as code (Terraform, Ansible), which is the same idea of a machine-readable model of the network.
Unit 5: Network management tools, SDN and AIOps
Tools and applications
| Category | Examples |
|---|---|
| Network monitoring | Nagios, Zabbix, LibreNMS, PRTG, SolarWinds, Prometheus with exporters |
| Visualisation | Grafana dashboards |
| Traffic analysis | Wireshark, tcpdump, ntopng, NetFlow collectors |
| Discovery and mapping | Network topology maps, nmap |
| Configuration management | Ansible, Terraform, Oxidized (config backup) |
| Logging | Syslog servers, ELK or Loki, SIEM |
| Simple tools | ping, traceroute, mtr, iperf |
Common measures: availability, latency, packet loss, jitter, bandwidth utilisation and error rates. Practices: baselining (know the normal), thresholds and alerting, capacity planning, root-cause analysis and change management.
Software-defined networking and virtualisation
SDN separates the control plane (decisions) from the data plane (forwarding): a central controller programs switches through an API such as OpenFlow. Network function virtualisation (NFV) runs firewalls, routers and load balancers as software. Result: networks are programmable, and management becomes software engineering. Cloud networking (VPCs, security groups, transit gateways) and Kubernetes networking (CNI plug-ins such as Cilium, Calico) are software-defined.
Observability and AIOps
| Idea | Meaning |
|---|---|
| Observability | Understand system state from metrics, logs and traces (the three pillars); OpenTelemetry standardises collection |
| Service-level objectives (SLOs) | Targets for user-facing behaviour, with error budgets |
| Alert quality | Alert on symptoms, avoid noise and alert fatigue |
| AIOps | Use machine learning on telemetry for anomaly detection, event correlation, noise reduction and automated remediation |
| Intent-based networking | Describe the desired outcome; the system configures and verifies it |
Today: an SNMP exporter for on-premises switches, Prometheus and Grafana for metrics, flow logs for traffic, OpenTelemetry for traces and an alerting pipeline with runbooks are a complete modern network management system. The FCAPS areas map to tools, and AIOps tools help correlate thousands of events during an incident. Familiarity with all of these is a direct advantage for a DevOps or SRE role.
One-minute recap
| Unit | Key idea | You meet it in |
|---|---|---|
| 1. Fundamentals | FCAPS, manager, agent, MIB, in-band and out-of-band | Monitoring programme design |
| 2. SNMP and MIB | Get, GetNext, Set, Trap; OIDs; community strings | snmpwalk, snmp_exporter, Zabbix |
| 3. SNMPv2, v3, RMON | GetBulk, 64-bit counters; USM and VACM; probes | SNMPv3 hardening, NetFlow and flow logs |
| 4. Telecom and broadband | TMN layers, OSS/BSS; NETCONF, YANG, streaming telemetry | Network automation, gNMI telemetry |
| 5. Tools, SDN, AIOps | Monitoring stacks, SDN, observability, AIOps | Prometheus and Grafana, OpenTelemetry, Cilium |
Next in Engineering Revisited: Pattern Recognition.
Written by Claude, Anthropic’s AI assistant, as part of Engineering Revisited, a quick refresher of the B.Tech subjects one at a time.