PracticeLabs
Week 6Lesson 5Required23 min estimated0% progress

Network Management Services: SNMP, Syslog, TFTP/FTP, QoS

Describe how SNMP monitors devices, how Syslog reports events by facility and severity, how TFTP and FTP move images and configs, and how QoS applies per-hop behaviour — then configure basic Syslog and a read-only SNMP community.

Lesson orientation

What you'll learn (5 objectives)~23 min: video → lesson → check → apply → lab prep

Learning objectives

  • Explain the SNMP manager/agent model and the difference between a get, a trap, and SNMPv2c versus SNMPv3 security
  • Order the eight Syslog severity levels and read a message's facility, severity, and mnemonic
  • Compare TFTP and FTP and use copy to back up or restore an IOS image or configuration
  • Describe the QoS per-hop behaviour pipeline and the EF and CS3 markings for voice
  • Configure timestamped Syslog to a server and a read-only SNMP community

Terms you will see

SNMP manager / agentOID / MIBGet / setTrap / informCommunity stringFacility / severity / mnemonicTFTP / FTPDSCP / CoSEF / CS3LLQPolicing / shaping

Time breakdown

  • Read the notes10 min
  • Walk the Syslog and SNMP configuration5 min
  • Practice in the Network Monitoring Console5 min
  • Predict the severity threshold3 min

Assigned video

Opens on YouTube
Recommended video

Free CCNA | SNMP | Day 40 | CCNA 200-301 Complete Course

By Jeremy's IT Lab · Opens externally on YouTube

Assigned watch
~6 min

Where to stop

Syslog, file transfer, and QoS are described in the written sections rather than in this video.

The source marks this segment timestampStatus="unresolved" with no start or end, so no bounds are shown here.

The link opens in a new browser tab. Return here when you finish watching.

Watch for these concepts

  • How a manager polls agents with get and how agents raise traps
  • Why SNMPv3's authentication and encryption beat v2c community strings

Go beyond the video

The assigned video introduces the idea. CCNA Practice Labs completes the learning path: clarify core concepts, explore how each device behaves, visualize the communication path, check your understanding, and prepare for hands-on lab work.

Watch the concept → understand it → visualize it → practice it → apply it.

SNMP — a manager, agents, and two directions of traffic

A central NMS (the manager) talks to an agent running on each device. Traffic flows both ways, and the distinction matters more than the details.

  • Manager to agent — a get reads a value (interface counters, CPU, uptime), identified by an OID within the device's MIB, the structured catalogue of readable objects. A set writes a value, which is rarer and more dangerous.
  • Agent to manager — a trap or inform is an unsolicited alert the agent raises when something happens, so the manager learns immediately rather than waiting for the next poll.
Versions, and why v3 is the answer to "which is secure"
VersionAuthenticationPrivacyVerdict
SNMPv2cPlaintext community string (e.g. public)None — the string crosses the wire in the clearRead-only is tolerable on a trusted segment; a read-write v2c community is a real exposure
SNMPv3Yes — verifies the senderYes — encrypts the payloadThe secure choice, and what production networks use
Trap versus inform
Both are agent-initiated alerts, and the difference is acknowledgement. A trap is fire-and-forget — the agent sends it and never learns whether it arrived. An inform is acknowledged by the manager, so the agent can retransmit if no acknowledgement comes back. Informs are more reliable and cost more overhead. For the exam: traps are unacknowledged, informs are acknowledged.

Syslog — eight severities, and lower is worse

Every Syslog message carries a facility (the subsystem that generated it), a severity level, and a mnemonic (a short name for the event). Read %LINK-3-UPDOWN: Interface Gi0/1, changed state to down as facility LINK, severity 3, mnemonic UPDOWN.

The eight severity levels — 0 is the most severe
LevelNameFeel
0EmergencySystem unusable
1AlertAct immediately
2CriticalCritical condition
3ErrorError condition
4WarningWarning condition
5NotificationNormal but notable
6InformationalRoutine info
7DebuggingDebug output
The scale runs backwards from instinct
Severity 0 is the most severe and 7 the least, which is the opposite of the "bigger number, bigger deal" instinct almost everyone brings. A threshold sends that level and every lower number, so logging trap informational (6) forwards levels 0 through 6 and drops only debugging. The mnemonic that survives exam pressure: Every Awesome Cisco Engineer Will Need Ice-cream Daily — Emergency, Alert, Critical, Error, Warning, Notification, Informational, Debugging.

Messages can go to several destinations at once: the console, the local logging buffer in RAM (read with show logging), a monitored VTY session via terminal monitor, or a central Syslog server for retention and correlation. That last one is exactly where NTP pays off — without synchronised clocks, timestamps from different devices cannot be lined up into a single story.

Moving files, and deciding who goes first

TFTP versus FTP
ProtocolTransportAuthenticationCharacter
TFTPUDP 69NoneTiny and simple; fine on a trusted LAN for quick config or image copies
FTPTCP 20/21Username and passwordReliable and authenticated; better for larger or less-trusted transfers

Both are driven from IOS with copy: copy running-config tftp backs the live config up to a server, and copy tftp flash pulls a new IOS image onto the device. TFTP's simplicity is also its weakness — no authentication, and UDP with no built-in reliability — so FTP, or secure SCP, is preferred wherever security or size matters.

QoS applies per-hop behaviour: how each device treats a packet as it passes through. When a link is congested, QoS decides whose traffic wins. The pipeline has four stages.

  1. Classify — identify traffic, by ACL, by an incoming marking, or by NBAR application recognition.
  2. Mark — stamp a value the rest of the path trusts: CoS (3 bits) in an 802.1Q tag at Layer 2, or DSCP (6 bits) in the IP header at Layer 3. Voice payload is marked EF (DSCP 46); voice signalling is CS3 (DSCP 24). Mark as close to the source as possible.
  3. Queue — during congestion, scheduling picks the dequeue order. LLQ (Low Latency Queuing) gives voice a strict priority queue and is the CCNA gold standard for real-time traffic.
  4. Police or shape — enforce a rate. Policing drops or re-marks the excess and is typically applied inbound; shaping buffers and delays the excess to smooth bursts and is typically applied outbound.
QoS does not create bandwidth
This is the mental correction most learners need. On a congested link QoS only reallocates what already exists, so prioritising voice necessarily deprioritises something else — there is no free capacity being conjured. Voice is chosen for priority because it is delay-, jitter-, and loss-sensitive in a way that bulk data simply is not: a file transfer that arrives a second late is fine, a voice packet that arrives a second late is useless.

Configure Syslog and a read-only community

Two of the four services get a short, exam-relevant configuration. TFTP and FTP are a copy command, and QoS per-hop behaviour is describe-only at CCNA level.

Timestamped Syslog to a server, plus a read-only SNMP community
! --- Syslog: timestamp messages and send Warning-and-worse to a server ---
R1(config)# service timestamps log datetime msec
R1(config)# logging host 10.10.20.15
R1(config)# logging trap warning

! --- SNMP: read-only monitoring for the NMS ---
R1(config)# snmp-server community CCNA-RO ro
R1(config)# snmp-server host 10.10.20.15 version 2c CCNA-RO

! --- Verify local logging ---
R1# show logging
Syslog logging: enabled
    Trap logging: level warnings, 42 message lines logged
    Logging to 10.10.20.15  (udp port 514, audit disabled)

! --- Back up the config off-box ---
R1# copy running-config tftp
  • service timestamps log datetime msec puts a real date and time on each message so server-side correlation across devices works — this is where the NTP lesson pays off.
  • logging trap warning sets the threshold to level 4, so levels 0 through 4 are sent and levels 5 through 7 are held back.
  • snmp-server community CCNA-RO ro defines a read-only community: the NMS can poll but change nothing. In production you would prefer SNMPv3 for authenticated, encrypted access rather than a plaintext community.
  • show logging confirms the destination and the threshold; Syslog rides UDP 514.

What you should retain

  • SNMP is manager/agent: get polls, trap and inform alert; v2c is a plaintext community, v3 adds authentication and encryption.
  • Syslog severity 0 to 7 with 0 Emergency most severe; each message carries a facility, severity, and mnemonic; a threshold sends that level and all lower numbers.
  • TFTP is UDP 69 with no authentication; FTP is TCP 20/21 and authenticated; both move images and configs with copy.
  • QoS per-hop behaviour: classify, mark (EF 46 voice, CS3 24 signalling), queue (LLQ for voice), then police (drop) or shape (buffer).
  • QoS reallocates bandwidth; it never adds any.
  • Config: logging host plus logging trap, service timestamps, and snmp-server community NAME ro.
Pause and predictNot scored — nothing is recorded

Before you read on

You set logging trap warning (level 4) toward the Syslog server. An interface-down event is generated at severity 3, and a debug message at severity 7. Which of the two reaches the server, and why?

Interactive tool
Optional~15 min

Network Monitoring Console

Complete the SNMP, Syslog, and NTP modes, then run Investigate the Flapping Port to correlate traps, syslog messages, and synchronized timestamps before finishing with the short practice check.

Study deeper

Topic guides extend this lesson — they do not replace the first-party walkthrough above.

QoS

For the full DSCP and CoS marking tables and the queuing algorithms behind LLQ