7. Path Conditioning.

mp4

Transcript
00:00:00

In this section we discuss path conditioning which includes forward error correction and packet order
correction.

00:00:07

Throughput in the network is negatively affected by latency and loss, and of course the speed of the
available links.

00:00:14

Latency and loss are inescapable facts of life for any network.

00:00:19

Typical land latencies are displayed above crossing the USA.

00:00:23

Expect to see 60 to 120 millisecond delays. International network connections can easily see up to 200
millisecond delays.

00:00:33

Satellite connections have the most latency with typical delays of about 1/2 second, with delays of
several seconds not unheard of.

00:00:41

If you buy bandwidth from a carrier, unless you have a private line, your service level agreement with
them will likely include provisions for allowable loss in the network.

00:00:51

An MPLS network typically has.

00:00:53

Boss, ranging from 0.1% to .5%. If your data traverses the public Internet.

00:01:01

Expect to see losses from 0.5% to 1% of course loss on the Internet can be much higher than this at peak
times or when there are changes in traffic loads for any number of reasons.

00:01:14
Packet loss and packets arriving at their destination out of order cause problems in your network in
many ways.

00:01:21

In addition to the slow response resulting from packets that need to be retransmitted, which might also
further congest an overloaded link, time sensitive applications like video and voice over IP can suffer
from quality and intelligibility problems caused by packets arriving out of order.

00:01:40

Most carriers.

00:01:41

Provided impl based networks have multiple paths with equal costs to any given destination.

00:01:47

And the different transmission times over each of the possible paths makes it possible that any two
given packets can arrive out of order.

00:01:55

The combination of loss and out of order packets can make video freeze and pixelate voiceover IP.

00:02:02

Audio quality can also suffer, resulting in pops and clicks jitter induced distortion.

00:02:08

And last or dropped calls in order to correct for the effects of lost packets in your network.

00:02:15

Edge Connect offers forward error correction or freq as part of its path conditioning technology.

00:02:22

**** works by creating extra packets containing parity data that it transmits, along with irregular data
packets within the second best primary link.

00:02:31

These extra parity packets can also be used by the receiving edge connect to reconstruct missing data it
have packet.

00:02:39

This lost in transit.

00:02:40

This is like the techniques.

Used in a raid storage array to rebuild lost data if, uh, redundant disk goes down in the array.

00:02:48

What this means is that you don't have to retransmit lost packets, making your network more.

00:02:53

Efficient, the ratio of fret packets to data packets can vary dynamically as network conditions change,
and edge connect appliances detect a packet loss in the network.

00:03:04

When there's no loss, very little bandwidth is used for freq.

00:03:09

There is some network overhead associated with the effect packets, as they essentially constitute extra
packets traversing the network.

00:03:16

This means if you're transmitting over a saturated Wan link, FAQ could make the problem worse by
trying to move more packets across a link that's already full.

00:03:27

It's important to know your network.

00:03:29

And make sure that you have adequate bandwidth available for your needs.

00:03:33

Higher vulnerability link bonding takes advantage of fecund parallel links to provide extremely high
reliability in.

00:03:41

This example we have an impl link and an Internet link that are logically bonded and on this link we
make use of 1 to one check one to one fact means that we send one freq parity packet for each regular
datastream packet.

00:03:56

In this example the Edge Connect on the left sends the regular data stream packets.

00:04:01

Over the link and 100% complete set of parity packets across the Internet link.

00:04:07
In transmission across the Wan, even though there could be substantial data loss in both links, the edge
connected the remote end can reconstruct the entire data stream from the parity packets.

00:04:19

You should never even notice the packet loss in transit because the devices at the remote end beyond
the remote edge connect never see it happen.

00:04:27

This is the kind of reliability you want for traffic that's sensitive to loss, like voice over IP and video
conferencing.

00:04:35

As we discussed earlier, packet switched to networks like MPLS Networks managed by service providers
almost certainly contain some out of order packet transmissions.

00:04:46

As we can see in this example, out of order, packets happen because even though packets might be
transmitted in the correct order, the individual packets might be taking different paths through the
network, subject to different latency and processing, which can result in them arriving at a destination
out of order.

00:05:04

This can cause needless retransmissions.

00:05:07

Because the receiving device perceives that the out of order packet has been lost and failed to
acknowledge that packet and subsequent packets causing the sender to have to retransmit edge
connect solves this problem by accurately measuring round trip time.

00:05:22

Knowing how long to wait for packets and cashing out of order packets, it can cache any arriving packets
temporarily while it waits for missing packets to arrive via a different path when they arrive, the local
edge connect transmits all the packets on the local land in.

00:05:39

The correct order.