Practice: addressing and media problems · CSCO-111 Lesson 12

It is the last week of November, and Dev has booked the Brightline meeting room for two hours on Friday afternoon. On the whiteboard he has written eight short problems taken from real tickets of the past three months, with the client names left in. Jordan's job is to solve each one out loud, explaining the reasoning, while Dev plays the part of an exam question that gives nothing away. Marisol sits in, partly to help and partly because she wants to see whether the flash cards worked.

This lesson brings together the whole course through practice. It starts with a method for reading exam-style questions, then works through problems on addressing, IPv6, Ethernet, media and wireless, each set in the clients you have met. Work each problem yourself before opening its solution. The lesson ends with a short list of what to review before moving on to CSCO-112.

Reading a question before answering it

Exam questions are short, and most of the information you need is in them. Read the whole question, including every option, before choosing. Underline, mentally or on the whiteboard you are given, the numbers and the key words: a prefix length, a port number, a distance, the word 'not', the word 'best'. Ask which layer the question is about; that often eliminates two options at once. For calculation questions, write the steps rather than doing them in your head, because a single slip with a block size or a power of two produces a wrong answer that is often listed as one of the options.

Watch for distractors, wrong options designed to look right. In subnetting questions, the most common distractors are the network address offered as a host address, the broadcast address of the neighbouring subnet, and a mask one bit off. In media questions, they are cables that work but not over the distance given, and connectors that fit the wrong kind of fibre. In wireless questions, they are overlapping 2.4 gigahertz channels and dBm values with the sign misread. Expect them and check each answer against the question before moving on.

Manage time. With an exam of around 50 minutes, you have roughly a minute per item if the commonly reported question counts are right. Answer what you know first, make your best choice on hard items, and come back if the exam allows it. Never leave a question blank when there is no penalty for guessing.

Problem types and the step that prevents the most common mistake.
Problem type	First step	Common trap
Find network or broadcast address	Find the interesting octet and block size	Using the mask value instead of 256 minus it
Count usable hosts	2 to the power of host bits, minus 2	Forgetting to subtract 2
Choose a prefix	Work from hosts needed plus growth	Choosing a subnet one bit too small
Same subnet?	Find each address's subnet	Assuming addresses close together are in the same subnet
Compress an IPv6 address	Drop leading zeros, then use one double colon	Using the double colon twice
Pick a cable	Check distance and speed together	Copper over 100 metres
Pick a Wi-Fi channel	Use 1, 6 and 11 at 2.4 GHz	Choosing overlapping channels

Problem set 1: IPv4 addressing

Worked problem

A ticket from Kettle & Crumb

A new tablet at the third shop was given the static address 192.168.40.127/26 by a manager following an old note. It cannot connect to anything. Why?

Show solution

A /26 has a block size of 64, so the subnets in the last octet are 0, 64, 128 and 192.
127 falls in the subnet 64 to 127.
127 is the last address in that block, which is the broadcast address of 192.168.40.64/26.
A host cannot use the broadcast address. A valid static address would be one of .65 to .126 outside the DHCP pool.

Answer: 192.168.40.127 is the broadcast address of its /26 subnet, so it cannot be assigned to a host.

Worked problem

A ticket from Prairie Paws

The second location's staff network is 10.20.10.0/24 with its gateway at 10.20.10.1. A new X-ray workstation was set to 10.20.10.70 with mask 255.255.255.192 and gateway 10.20.10.1. It reaches some staff PCs but not others, and nothing beyond the site. What is wrong?

Show solution

With a /26 mask, the workstation thinks its network is 10.20.10.64 to 10.20.10.127.
Staff PCs between .65 and .126 appear local, so it sends to them directly and they work.
Everything else, including PCs below .64, appears remote, so it must use its gateway.
Its gateway, 10.20.10.1, is outside what it thinks is its own network, so it cannot reach the gateway at all.
That explains both symptoms. The mask should be 255.255.255.0 to match the /24 network.

Answer: The subnet mask is wrong. With /26 the gateway appears to be on another network; set the mask to 255.255.255.0.

Worked problem

Planning an addition

Northfield's new computer lab needs its own subnet for 22 PCs, a printer, a teacher's laptop and an access point, with room for about 50 more devices if the lab doubles. The centre has 10.40.0.0/16 for all its networks. Which prefix should the lab get, and what is a suitable first subnet if 10.40.5.0/24 is free?

Show solution

Today: 22 + 1 + 1 + 1 = 25 devices; after growth about 75.
6 host bits give 62 hosts, too few for 75; 7 host bits give 126.
The prefix is 32 − 7 = /25.
10.40.5.0/25 covers 10.40.5.0 to 10.40.5.127, with hosts .1 to .126 and the gateway at .1.

Answer: A /25, such as 10.40.5.0/25, with hosts from 10.40.5.1 to 10.40.5.126.

A small generator of subnetting practice. Change the seed for a new set.

import ipaddress, random

random.seed(111)
print('Practice set (answers below each question):')
for n in range(1, 6):
    prefix = random.choice([25, 26, 27, 28, 29, 30])
    host = ipaddress.ip_address(random.choice(['192.168', '172.20', '10.4']) + f'.{random.randint(0, 255)}.{random.randint(1, 254)}')
    net = ipaddress.ip_interface(f'{host}/{prefix}').network
    print(f'{n}. {host}/{prefix}: network? broadcast? usable hosts?')
    print(f'   -> {net.network_address}, {net.broadcast_address}, {net.num_addresses - 2}')

Output

Practice set (answers below each question):
1. 172.20.252.50/26: network? broadcast? usable hosts?
   -> 172.20.252.0, 172.20.252.63, 62
2. 172.20.86.162/28: network? broadcast? usable hosts?
   -> 172.20.86.160, 172.20.86.175, 14
3. 10.4.215.144/26: network? broadcast? usable hosts?
   -> 10.4.215.128, 10.4.215.191, 62
4. 192.168.237.107/26: network? broadcast? usable hosts?
   -> 192.168.237.64, 192.168.237.127, 62
5. 10.4.142.97/28: network? broadcast? usable hosts?
   -> 10.4.142.96, 10.4.142.111, 14

Try it yourself

Describe a subnet

Write summary(cidr) that takes an address with prefix, such as '192.168.40.150/27', and returns a tuple of four strings and one number: network address, broadcast address, first usable host, last usable host, and the count of usable hosts. You may use the ipaddress module. Assume prefixes from /8 to /30.

import ipaddress

def summary(cidr):
    pass

print(summary('192.168.40.150/27'))

Problem set 2: IPv6 and layer 2

Worked problem

Reading an IPv6 address

A printer at Riverside reports the address fd4a:1b2c:9e00:20::1c/64. What type of address is it, what is its network prefix, and what is its full form?

Show solution

The address begins fd, so it is a unique local address, for internal use only.
With a /64, the prefix is the first four groups: fd4a:1b2c:9e00:20::/64.
Four groups appear before the double colon and one after, so it stands for three zero groups.
Full form: fd4a:1b2c:9e00:0020:0000:0000:0000:001c.

Answer: A unique local address on fd4a:1b2c:9e00:20::/64; full form fd4a:1b2c:9e00:0020:0000:0000:0000:001c.

Worked problem

Frames and MAC addresses

A switch at Prairie Paws shows the same MAC address, 0011.3245.8f20, learned first on port 5 and then on port 12, flapping back and forth every few seconds. Port 12 connects to a small unmanaged switch in an exam room. What are two likely causes?

Show solution

A MAC address should normally be learned on one port at a time.
One cause is a duplicate MAC address, such as a cloned address on another device, like the Kettle & Crumb tablet.
Another is a loop: if the exam-room switch is also connected back to the main switch by a second cable, frames circle and the switch sees the same address arriving on two ports.
Check the exam-room switch for a second uplink cable first, because loops can cause broadcast storms that affect the whole network.

Answer: A duplicate MAC address or, more urgently, a loop through the unmanaged switch.

Problem set 3: media and wireless

Worked problem

Choosing media

Kettle & Crumb's office is moving to the second floor of a building where the provider's fibre enters in the basement, 140 metres of cable route away. The office needs 1 gigabit now and possibly 10 gigabit later. What should connect the basement to the office?

Show solution

140 metres is beyond the 100-metre limit for twisted-pair Ethernet.
Fibre is needed. Over 140 metres inside a building, multimode fibre is enough: OM4 supports 10 gigabit to about 400 metres with 10GBASE-SR.
Singlemode would also work and allows more growth, but transceivers for it have traditionally cost more.
Use SFP or SFP+ modules that match the fibre at each end, and a riser- or plenum-rated cable as the route requires.

Answer: A multimode OM4 fibre pair (or singlemode) with matching transceivers; copper cannot cover 140 metres.

Worked problem

Checking a PoE plan

The third shop's 8-port PoE switch has a 120-watt budget. The planned devices are two Wi-Fi 6 access points at 802.3at, which the switch reserves at 30 watts each, four cameras at Class 3, reserved at 15.4 watts each, and a phone at Class 2, reserved at 7 watts. Is the budget enough?

Show solution

Access points: 2 × 30 = 60 watts.
Cameras: 4 × 15.4 = 61.6 watts.
Phone: 7 watts.
Total: 60 + 61.6 + 7 = 128.6 watts, which is more than 120.
Options include a switch with a larger budget, moving the phone to an injector, or setting the switch to allocate by actual draw if the model supports it.

Answer: No. The plan needs 128.6 watts against a 120-watt budget.

Worked problem

Fixing a crowded band

Riverside's office has three access points, all on 2.4 gigahertz channel 11 and 5 gigahertz channel 36 at 80 megahertz. Lawyers complain of slow video calls. What changes would help?

Show solution

Three access points on the same channel must share airtime with each other.
At 2.4 gigahertz, spread them across channels 1, 6 and 11.
At 5 gigahertz, give each access point a different channel; narrowing from 80 to 40 megahertz makes more separate channels available.
Check signal strength at the lawyers' desks, aiming for about −67 dBm or better for video.

Answer: Put the access points on different channels in both bands, narrow the 5 gigahertz channels, and check signal levels.

Problem set 4: ports and topology

Worked problem

A blocked application

After a firewall change, Prairie Paws' staff can browse websites but cannot send email from their desktop mail program, although webmail works. The program is set to submit mail to the server on port 587. What should the technician check?

Show solution

Browsing and webmail use HTTPS on TCP 443, which is clearly allowed.
The mail program submits mail with SMTP on TCP 587.
The firewall change probably removed the rule allowing outbound TCP 587 to the mail provider.
Restore the rule, test, and record the change in the ticket.

Answer: Check for a missing outbound rule for TCP 587, SMTP submission, on the firewall.

Worked problem

Choosing a WAN shape

Kettle & Crumb's point-of-sale system runs from a server at the office. Each of the three shops needs to reach the office, but shops never talk to each other. Which design fits, and how many tunnels are needed?

Show solution

Traffic flows only between each shop and the office.
A hub-and-spoke design with the office as the hub matches that flow.
Each shop needs one tunnel to the office: 3 tunnels.
A full mesh of four sites would need 6 tunnels, adding cost and complexity without benefit.

Answer: Hub-and-spoke with the office as hub, using 3 tunnels.

What to review before CSCO-112

By the end of the session, Jordan has solved seven of the eight problems without help and the eighth, the PoE budget, after one hint. Dev's advice is the same for every learner finishing this course. Practise subnetting until you can find the network, broadcast and host range of any /24 to /30 address in under a minute, then extend to /16 to /23. Write out the OSI layers and the PDU names from memory. Recite the common ports in groups. Draw the T568B colour order and the 2.4 gigahertz channel plan. Compress and expand IPv6 addresses until the rules are automatic. Then begin CSCO-112, which uses everything here to configure and troubleshoot real devices.

A self-check before moving on.
Topic	Lesson	Ready when you can
OSI and TCP/IP models	2	List the layers, PDUs and example protocols from memory
Ethernet and MAC addresses	3	Read a MAC address, explain switch learning and ARP
IPv4 and masks	4	Convert masks and prefixes, recognize special ranges
Subnetting	5	Solve any /24 to /30 problem in under a minute
IPv6	6	Compress, expand and identify address types
Cabling	7	Choose copper or fibre by distance and speed; check PoE
Wireless	8	Pick standards, bands and non-overlapping channels
Devices	9	Name devices and their layers in any network cupboard
Ports	10	Match common ports and say TCP or UDP
Topologies and diagrams	11	Count mesh links and choose the right diagram

Marisol, collecting the flash cards at the end of the session, says Jordan is ready for the next course. Dev agrees, then adds one more problem to the whiteboard for next week: the Prairie Paws network that started this course, drawn from memory, with every address, port and cable labelled.

Common mistakes

MistakeAnswering before reading every option.
InsteadRead the whole question and all options; distractors are designed to catch quick answers.
MistakeDoing subnet arithmetic in your head under time pressure.
InsteadWrite the block size and boundaries on the whiteboard; it takes seconds and prevents slips.
MistakeTreating a working connection as proof the design is right.
InsteadCheck limits such as distance, PoE budget and channel overlap even when things seem to work.
MistakeLeaving questions blank.
InsteadMake your best choice; you can come back if time allows.
MistakeStudying only the topics you enjoy.
InsteadCisco does not publish domain weights, so review every area.

Summary and key terms

Read the whole question, find the layer and the key numbers, and watch for distractors.
Write out subnet steps: interesting octet, block size, boundaries, hosts.
Broadcast addresses cannot be host addresses, and masks must match the network.
IPv6 types are recognized by their first digits; expand double colons by counting groups.
Choose media by distance and speed together, and check PoE budgets and Wi-Fi channels.
Ports, topologies and diagrams turn symptoms into likely causes.
Review every lesson's skills before moving to CSCO-112.

Distractor: A wrong answer option designed to look correct.
Block size: The size of each subnet in the interesting octet, 256 minus the mask value.
Broadcast address: The last address in a subnet, which cannot be assigned to a host.
Loop: A layer 2 path that lets frames circle endlessly, causing MAC flapping and broadcast storms.
Power budget: The total PoE power a switch can supply to all its ports.
Hub and spoke: A WAN design where branch sites connect only to a central site.

Practice problems

Work each problem on paper first, then reveal the answer to check.

Is 10.1.1.63/26 a usable host address?
Show answer
No; it is the broadcast address of 10.1.1.0/26.
Which prefix fits 100 hosts with the fewest addresses?
Show answer
/25.
What does a MAC address flapping between two ports suggest?
Show answer
A duplicate MAC address or a loop.
Can Cat 6A carry 10 gigabit Ethernet 140 metres?
Show answer
No; twisted pair is limited to 100 metres.
What is the full form of 2001:db8::1?
Show answer
2001:0db8:0000:0000:0000:0000:0000:0001.
Which port does SMTP submission from a mail client usually use?
Show answer
TCP 587.
How many tunnels does hub-and-spoke need for one hub and five branches?
Show answer
Five.