This video reviews key areas of Layer 2 Network Mapping.

If you have any questions, please contact us at SALES@UVEXPLORER.COM or +1 (801) 901-0825.

Download UVexplorer Now:
[ Ссылка ]

Transcript:
Thank you for watching this video on understanding
layer 2 networking.
Today we will review the basics of Ethernet traffic
and how it works.The mechanisms behind it and review cam tables.
So the cam tables are the fundamental operation of
what drives all of these network switches.
And everything's based on Mac address based
communication.
Mac addresses are basically like your home address.
In this particular case, H one computer has an address
of A A A for a Mac address, two has an address of B B
B, and H three has a Mac address of C C C. These are
obviously simplified for this particular example so
that we can move through it quickly and make it a
little bit easier to digest.
But when computers first started doing networks,
everything was based off of something
called C.s.m.a. C.D.
which is an abbreviation for carrier sensing multiple
access with collision detection.
And that's a fancy way of saying that when I send a
packet out across the network, I'm going to send it to
all ports or flood it to the destination.
I rely on the honor system to determine that you are
the destination header and should receive the packet.
In today's environment, this does not work from a
perspective of volume of traffic but there are serious
cybersecurity initiatives that would be violated by
sending your traffic flows out to all ports.
So everything on Ethernet switching today is based on
something called the flood and learn concept.
So in this example we're trying to send an Ethernet
frame to a destination of Mac address BBB with a
source address of AAA.
So in other words, we're sending this
from h one to h two.When this goes out across the network, the cam table
that holds all the mappings for the Mac address ports
initially comes back as being 100% empty.
When the first packet gets sent out across the
network, it executes the concept of flood and learn.
Now we can see that for H one, my own Mac address our
source is aaa.
While I don't know where the destination
Mac is for BBB,
so when that packet goes out across the network, we
flood that packet out to all ports.
During this process, I can determine my own Mac
address and record that in the table.
As each entity starts to send things across the
network, you actually build this Mac address table.
And once you learn all the Mac addresses for all the
ports, the switch no longer operates in a flood
environment.
It's learned the pathway and will do a directed
pathway of delivery.
Today, all switches work on this concept
of flood and learn.
When we talk about segmenting the network down, you
start talking about how large that domain is, that
you're actually going to be doing that
flood and learn process.
So what can go wrong in this type of environment?
Now, there's something called bum traffic, and while
that seems like an absolutely ridiculous acronym to
put out there, bum traffic stands for broadcast,
unknown and multicast.
Those are the three types of traffic that will go
across your network that cannot be commonly mapped
back to a cam table, and their only representative
action at that point is to flood those
out to all ports.

Now, multicast, unfortunately, has gotten a bad rep in
the industry as being problematic or
difficult to deal with.
A lot of people who've dealt with multicast
transmissions before have seen instances where they
get complete network failure.
Multicast only works in properly designed and
implemented networks.
Multicast is the most efficient way to deliver high
volume data across the spectrum.
So in this example to reach the three endpoints with
unicast, I'm sending three different
versions of that traffic.
Where on the multicast, I'm sending one individual
entity, and through group subscriptions or I.G.N.P.
subscription, it's being delivered to the actual
monitors that are requesting it without
increasing the bandwidth.
Unfortunately, if you don't do unicast or multicast,
you get into the third scenario, which is flooding
data across your entire network.
These three types of traffic are always problematic
for certain high volume data transmission systems,
like video surveillance, and it's important to know
when these particular events are happening.
So we talk about network segmentation, there are some
vendors that believe in the proxy approach.
They'll take an archiver, they'll put one network card
in it for the high volume data to come into, and one
network card for it to send out, and put them on two
separate networks.
And then they rely on that particular archiver to
manage that traffic.

свернуть