If I interview someone for a job, and they claim CCNA level skills – then I’m going to expect them to know the topics within the scope of CCNA pretty well. When Cisco spins out a question set that you get for the CCNA exam(640-802,640-760,640-553……..), I imagine they want to assess whether you know the topic to a similar depth/breadth. My theory then is that a job interview that focuses on CCNA level topics is a pretty good tool to assess whether you will do well on the CCNA certification exam. A nice theory, but to put something tangible into the discussion, today I’ll look deeper at the responses I’d hope to hear from the candidate. You can gauge for yourself whether you think that a fake interview like the one contained here, with your CCNP co-worker, would help you decide when to take the plunge.
Spoiler alert: if you wanted to try to answer for yourself, and haven’t read last week’s post yet, stop, look here first, and then come back.
What I’m searching for is whether the candidate knows how a small internetwork works under normal circumstances. In my opinion, you can’t troubleshoot very well if you don’t understand how something works normally. And that understanding has to be pretty thorough. And while the CCNA exam does test some troubleshooting skills, I’d be ok with a “CCNA level job candidate” if they just knew some instances of typical problems, but not all.
Here’s what I’d hope to hear in the interview to start. At this stage, I’d ask for a general description, as mentioned back in the previous blog post.
Bring up router-router links:
- On the HDLC link, I’d like to hear about the line coming up after the no shut, the need for keepalives to flow each direction before the line protocol would come up.
- On the PPP link (R2-R3), I’d like to hear that PPP LCP was in control of the keepalive process. I’d also want to hear that IPCP and CDPCP would need to complete before IP and CDP traffic could pass the link.
- For both, I’d want tem to know that CDP would be working on both links.
- That RIP begins to send messages and exchange routes.
Bring up router-switch links:
- Describe the 802.3u autonegotiation process, and why each link ends up with 100/full.
- Note that CDP runs over the link.
- Note that no trunking is used on these links.
- That RIP now advertises the LAN subnets, whereas before these links came up, the routers did not advert the associated subnets.
Bring up switch-PC links:
- Autonegotiation details again.
- That the switch sends CDP on the links, but hears nothing in return.
- Host PC1 asks for an address, mask, default gateway, and DNS IP addresses from the DHCP server.
- The other two hosts, noted as having static IP addresses, do not ask for DHCP-leased addresses.
PC1 webs to PC2:
- PC1 ARPs to find it’s default gateway’s MAC address (172.16.1.1, R1-MAC for short)
- PC1 sends a DNS request for hostname PC2.certskills.com to the DNS, which sends back a reply of 192.168.1.2.
- PC1 sends the first IP packet to 192.168.1.2, then the packet is routed.
- That no ARP is required on the HDLC or PPP links.
- Before R3 sends the packet over the LAN, R3 ARPs to learn PC2′s MAC address. (I’m assuming no other traffic has occurred other than what’s discussed here, but in real life, I’m sure the server would have already sent/received something, driving the ARP process.)
- Packets flow to establish the TCP connection (3-way handshake) between PC1 and PC2.
I’d probably stop them at that point, and ask clarifying questions. What would you want to ask if you were holding the interview? Where would you drill down? What did I leave out that you’d want to ask about? (I left out a few things on purpose to give you a small puzzle; I’ll post a comment in a few days to fill in the holes if no one comments/asks before then.)
For this job candidate – and the self-assessing CCNA candidate – I’d expect the above list to flow pretty easily, at least from the point of having looking at the topo and having looked at the config. The above is a relatively high level. If you’re prepping for CCNA, and these topics don’t come to mind when you see a small internetwork like the one in this example, then I’d expect that the troubleshooting required for some Sim and Simlet questions will be hit-or-miss. Of course, some are less important than others – the keepalive stuff may seem like trivia – but I’d want these candidates to get most of the above without a lot of struggle.
The next wave is when this gets fun. Call it 10 minutes to look at the topo, the config, and make the above list (briefly) in conversational style. In the next part of the interview, I might drill down on the following:
-
- Exactly what routes will R3 advertise to R2 when only the WAN links are up? Once the LAN links are up?
- What messages will flow between PC1 and the DHCP server?
- What source/destination IP addresses will exist in the DHCP messages, both while on PC1′s LAN and on the other side of R1?
- If R1′s Fa0/0 were configured for 802.1Q, what would need to be configured on the switch?
- If you could login only to R1 and R2, and you didn’t know the topology, what parts of the topo would you expect to be able to re-create?
- If you configured R2′s S0/0/1 with an IP address in a different subnet than R1′s S0/0/0, what would be the effect? Same question, but on the R2/R3 link?
- What routes would you expect to see in the IP routing table before any of these presumed changes happen?
- What entries would you expect to see in the MAC address tables of the three switches?
- What ARP table entries would you expect to see in R1? R2? R3?
I’d suggest that a candidate who could answer most of this with relative confidence shows the signs that they “get it”, that they are mastering some of the most important basics. I’d probably want a broader discussion, toss in some Frame Relay, multiple switches in a LAN for trunking and STP, but that’s a pretty good start.
For those of you working on CCNA out there, if you’re into this, tell me your answers to the questions above, and maybe a confidence factor. I’ll chime in to answer further for any of these that you want to ask about.
