Keeping a network running: monitoring, availability, performance, change control and documentation. Management protocols are in Network Management Protocols.
Monitoring
Metrics
| Metric | Meaning |
|---|---|
| Bandwidth | Maximum capacity |
| Throughput | Actual data rate achieved |
| Latency | Delay (ms) |
| Jitter | Variation in latency, matters for voice / video |
| Packet loss | % of packets not delivered |
| Utilisation | % of capacity used |
| Errors / discards | CRC, drops, see Network Troubleshooting |
| Availability | Uptime %, e.g. 99.99% ≈ 52 min downtime / year |
A baseline (normal performance) lets you spot anomalies.
Protocols and tools
| Tool / protocol | Purpose |
|---|---|
| SNMP (UDP 161, traps 162) | Poll devices (OIDs in a MIB); v1/v2c community strings (weak), v3 auth + encryption. Components: manager, agent, MIB |
| Syslog (UDP 514) | Central logs, severities 0 Emergency to 7 Debug |
| NetFlow / IPFIX / sFlow | Flow records: who talks to whom, how much |
| Packet capture | Wireshark / tcpdump deep dive |
| Ping / traceroute | Reachability and path |
| WMI / agents | Host metrics |
| Nagios, Zabbix, PRTG, LibreNMS, SolarWinds | Monitoring platforms |
| Prometheus + Grafana | Metrics and dashboards (common with Kubernetes) |
| SIEM | Correlate logs for security |
Syslog severities: 0 Emergency, 1 Alert, 2 Critical, 3 Error, 4 Warning, 5 Notice, 6 Informational, 7 Debug.
$ snmpwalk -v2c -c public 192.168.1.1 system
$ snmpget -v3 -l authPriv -u monitor -a SHA -A authpass -x AES -X privpass 192.168.1.1 sysUpTime.0
Cloud equivalents: 6. Azure Governance and Cost Management (Azure Monitor), 6. AWS Developer Services and Serverless (CloudWatch), VPC flow logs (5. AWS Networking).
High Availability and Redundancy
Concepts
- Single point of failure (SPOF): remove with redundancy
- Active-active (both share load) vs active-passive / standby (failover)
- MTBF, MTTR, MTTF; RTO (time to restore) and RPO (data loss window)
Techniques
| Technique | Purpose |
|---|---|
| Redundant links and devices | Survive failure of a path or box |
| Link aggregation (LACP) | Bandwidth + redundancy (Switching and VLANs) |
| STP / RSTP | Loop-free redundant switching |
| FHRP: HSRP, VRRP, GLBP | Shared virtual gateway IP so hosts keep one default gateway when a router dies |
| Dynamic routing | Reroute automatically (Routing Protocols) |
| Multihoming / ISP redundancy | Two providers, often with BGP |
| Load balancing / clustering | Share load, survive node loss |
| UPS and generators | Power continuity |
| NIC teaming | Server-side redundancy |
| RAID | Disk redundancy |
Site recovery types
- Hot site: fully equipped, live data, minutes to switch
- Warm site: equipment ready, data restored first, hours to days
- Cold site: empty space and power, days to weeks
- Cloud DR: see 7. AWS Solutions Architecture
Backups
- Full, incremental (since last backup, small, slow restore), differential (since last full)
- 3-2-1 rule: 3 copies, 2 media, 1 offsite; test restores
Quality of Service (QoS)
Prioritises traffic when links are congested so latency-sensitive flows (voice, video) win over bulk transfers.
- Classification and marking: DSCP (IP header), CoS / 802.1p (layer 2). Voice often EF (46)
- Queuing: FIFO, priority queuing, weighted fair queuing, LLQ (strict priority for voice)
- Traffic shaping (buffer and smooth) vs policing (drop / re-mark excess)
- Congestion avoidance: RED / WRED
- Voice targets: one-way latency < 150 ms, jitter < 30 ms, loss < 1%
Traffic and Performance Concepts
- Unicast, multicast (one-to-group, IGMP) and broadcast; anycast
- Bandwidth management, traffic shaping, caching and CDNs
- MTU (1500 for Ethernet) and MSS; fragmentation hurts performance
- Duplex: full vs half; auto-negotiation
- 802.3 standards: 10/100/1000BASE-T copper to 100 m; 10GBASE-T; fibre SX/LX/LR for multimode / single-mode (Ethernet, Cabling)
Change and Configuration Management
- Change management: request → risk assessment → approval → schedule window → implement → verify → rollback plan → document
- Configuration management: baselines, version control for device configs (Git), config backups, golden images, templates
- Network as code / automation: Ansible, Python (Netmiko, NAPALM), Terraform; APIs (REST, NETCONF / RESTCONF, gNMI), SDN (control plane separated from data plane), intent-based networking
- Patch and firmware management: test, stage, deploy, verify
- Lifecycle: end-of-life / end-of-support planning
Documentation and Policies
| Document | Content |
|---|---|
| Physical diagram | Racks, cabling, ports, patch panels (MDF / IDF) |
| Logical diagram | Subnets, VLANs, routing, flows |
| IP address management (IPAM) | Allocations, DHCP scopes |
| Wiring / port maps | Cable and patch panel labels |
| Asset inventory | Devices, serials, warranties, firmware |
| Baseline / SLA | Expected performance and uptime commitments |
| Policies | Acceptable use, BYOD, remote access, password, incident response, DR / BCP |
Good practice: label everything, keep documentation current, use standard operating procedures, lock and log physical access.
Related Technology Notes
- IPv6 transition: dual stack, tunnelling, NAT64 (Ipv4 and Ipv6)
- Virtualisation / cloud networking: virtual switches, VNets / VPCs, NFV, 3. Azure Networking, 5. AWS Networking
- Wireless operations: site surveys, channel planning, controller management (Wifi)
- Voice / video: SIP, RTP, QoS
- IoT and OT networks: segmentation, specialised protocols