Device Management Access
A switch you cannot reach is a brick with lights. Managing a Cisco device means choosing where to reach it (console or network) and how to authenticate who does — Telnet, SSH, HTTP, or a central AAA server. This lesson wires the two ideas together: in-band versus out-of-band access, VTY lines, SSH setup, and where TACACS+ and RADIUS sit.
- Distinguish in-band from out-of-band device management.
- Describe the console port and its role as the zero-configuration recovery path.
- Configure VTY access with line vty 0 15 and restrict transport with transport input ssh.
- Prepare SSH with ip domain-name and crypto key generate rsa.
- State the ports: Telnet TCP 23, SSH TCP 22, HTTP 80 / HTTPS 443.
- Compare AAA with TACACS+ (TCP 49) and RADIUS (UDP 1812/1813), and mention cloud-managed management.
In-band vs out-of-band
Management access comes in two flavours, and the distinction is where the path travels:
- In-band — the management traffic shares the same network path as user data. SSH/Telnet/HTTP to a switch’s management (SVI) address travels across the infrastructure being managed. Convenient, but if the network is down, so is your ability to fix it.
- Out-of-band — the management path is physically separate, usually a console cable or a dedicated management LAN/VLAN with its own links. It works even when the data plane does not, which is exactly when you need it.
Best practice for the exam: the console is your out-of-band lifeline; everything else is in-band until you build a dedicated management network. The sentence you can quote when a question asks “why console access after a total network failure?” is: a console session does not depend on IP connectivity.
The console port
The console port is the first thing every Cisco device boots with: a serial management interface giving you a CLI before any networking is configured. It needs a console cable (rolled RJ-45-to-RJ-45, or USB) and a terminal emulator at 9600 baud, 8 data bits, no parity, 1 stop bit. Router-on-first-boot questions on the CCNA normally demand that you begin over the console, then enable remote access so you never need it again for routine work.
Because console access bypasses all IP controls, it is protected in IOS by its own line authentication:
R1(config)# line console 0 R1(config-line)# password cisco R1(config-line)# login R1(config-line)# exit
VTY lines and remote access
Remote management terminals on a Cisco device are VTY lines (Virtual TeleTYpe). Classic routers have lines 0–4; modern IOS switches and routers have 0–15. Each VTY is a slot for one incoming session, and the exam asks which protocols each line accepts and how to secure it:
R1(config)# line vty 0 15 R1(config-line)# password cisco R1(config-line)# login R1(config-line)# transport input ssh ! allow SSH only — no Telnet R1(config-line)# exit
transport input ssh is the secure lockdown: only SSH sessions may come in on these
lines. transport input all allows both SSH and Telnet; transport input none
forbids remote access entirely (console only). A configuration that answers a security question with
“prevent Telnet management” will contain exactly this line. The complementary
choice, transport input telnet, is the one the exam treats as the insecure legacy
default.
ip domain-name <name> followed by crypto key generate rsa —
without a domain name and RSA keypair, SSH cannot start.
Telnet vs SSH, and SSH setup
Telnet transmits everything — usernames, passwords, commands — in clear text. Any sniffer on the path reads it verbatim, which is why Telnet is legacy and unacceptable on a production network. SSH encrypts the whole session, authenticates the server with public-key cryptography, and is the standard management channel. Enabling SSH on IOS needs a small, exact sequence:
R1(config)# ip domain-name example.com R1(config)# crypto key generate rsa The name for the keys will be: R1.example.com Choose the size of the key modulus ... : 2048 R1(config)# username admin privilege 15 secret cisco R1(config)# line vty 0 15 R1(config-line)# login local R1(config-line)# transport input ssh R1(config-line)# exit
Three dependencies matter: the hostname must be set (the key name is built from
hostname.domainname), the RSA keypair must exist (the crypto key generate rsa step), and
the VTY lines must both authenticate and accept SSH. The login local directive tells the
device to check the local username database instead of a shared password. Then a client reaches it as
ssh admin@192.168.1.2.
crypto key generate rsa work
cleanly. If a question lists the steps out of order, order is the trap — domain first, key
second.
HTTP/HTTPS GUI management
Cisco devices also expose a web management GUI. It is enabled with ip http server (HTTP, port 80) and ip http secure-server (HTTPS, port 443). The GUI is a convenience, but the syllabus always hands you the negative view too: plain HTTP is cleartext, so HTTPS is the only acceptable web management, and in locked-down designs even that is switched off in favour of CLI-only access. When a security question offers “disable HTTP management on the switch”, the answer is no ip http server (and usually no ip http secure-server if HTTPS is unwanted too).
AAA: TACACS+ vs RADIUS
Local users live on the device; on a network with many devices, users and credentials should live on one server. AAA (Authentication, Authorization, and Accounting) is the framework, and the two servers you must contrast are TACACS+ and RADIUS:
| Characteristic | TACACS+ | RADIUS |
|---|---|---|
| Transport | TCP | UDP |
| Port | 49 | 1812 (auth) / 1813 (accounting) |
| Encryption | Encrypts the whole packet (headers included) | Encrypts only the password (authenticator) |
| AAA separation | Authentication, authorization, and accounting kept separate | Authorization merged with authentication (attributes returned with the access-accept) |
| Origin | Cisco-proprietary (open spec) | IETF standard (RFC 2865/2866) |
The canonical enabling command is aaa new-model, which activates AAA globally; after
that, methods such as aaa authentication login default group tacacs+ local chain the
server first and the local database as backup. The exam-grade facts are the transport and port pair:
TACACS+ = TCP 49, RADIUS = UDP 1812/1813 — both canonical in the blueprint.
Cloud-managed management
Modern platforms increasingly split management out of the on-premises CLI. Meraki dashboards and Catalyst Center/DNA Center manage devices from the cloud or a central controller, applying templates, monitoring, and assurance on your behalf. The CCNA’s interest is conceptual: cloud-managed devices still need a bootstrap connection (often at the console) but the ongoing administration happens through a controller and southbound APIs, which is the automation story Weeks 14–15 pick up. The phrase to keep is “management plane” — the plane you are interacting with in every command in this lesson, separate from the forwarding plane beneath it.
transport input ssh and login local on VTY lines, HTTPS rather than HTTP
if the GUI must run, and AAA through TACACS+ or RADIUS once the network is big enough. That sentence
is a complete “secure device access” answer in one breath.
The vocabulary you need
Every term below appears in the lesson text above. Select one to read its definition, then use the quiz at the end to check what stuck.
Review
Check what you have learned with these review questions.
Sources & references
- Cisco. CCNA (200-301) Exam Topics. cisco.com.
- IETF. Request for Comments (RFCs). rfc-editor.org.
- IANA. Internet Assigned Numbers Authority registries. iana.org.
Target exams: CCNA 200-301 v1.1 · CyberOps Associate 200-201. Last content verification: 2026-10-11.