Just to prove that he really knows what he is talking about, one of our field network Engineers decided to compose some ‘handy reference material’ for our more technically minded readers to help clarify what all this means:
Let’s first look at Latency
Latency is the time between a source sending a data packet and the destination receiving it. Round trip latency is more often quoted and this is the time for a request to be sent and the reply to be received.
Data travels at the speed of light, approximately three million meters / second.
The circumference of the world is approximately 40,000km, so in perfect conditions a data packet would take 100ms to circumnavigate the globe. This is the baseline speed for data speed. However, with each network hop the data packet has to go through a switch, so the packet is delayed by the decision making that the switch has to perform to send the packet to the correct next hop. This all adds to the latency.
Obviously, the nearer the destination is physically, the fewer switches a data packet has to go through, so the better the latency.
The effect that a network’s latency has depends upon the type of application. Any system that requires full two-way communication, for instance a video conference or VoIP call, would suffer from a high latency. With one-way communication, for instance email or the streaming of media, network latency is less of an issue.
According to leading IT company Cisco, for applications that require a very low latency a reading of 150ms is ‘acceptable’, that is to say a ‘return trip time’ or RTT of 300ms.
Network latency also depends upon the type of network technology being used:
- Local Ethernet — Up to 1ms
- Leased Line — 2-4ms, depending on length
- VDSL/Cable — 15-25ms
- ISDN — 25-45ms
- 56k Modem — 150-300ms
- 3G/4G GSM — 300-900ms
- Satellite — >700ms, mostly due to the distance from earth to the satellite in space (36,000km away)
Jitter is the difference in the latency, in milliseconds, of a packet flow between two systems, when some packets take longer to travel from one system to the other.
Pinging amazon.com.au [18.104.22.168] with 32 bytes of data four times gives different jitter results:
- Reply from 22.214.171.124: bytes=32 time=184ms TTL=237
- Reply from 126.96.36.199: bytes=32 time=197ms TTL=237
- Reply from 188.8.131.52: bytes=32 time=194ms TTL=237
- Reply from 184.108.40.206: bytes=32 time=175ms TTL=237
The first of these becomes our base line:
- Jitter base line
- Jitter between 1 and 2: 13ms
- Jitter between 2 and 3: -3ms
- Jitter between 3 and 4: -19ms
Jitter results from network congestion, timing drift and route changes.
With high jitter you could have three packets not sent when requested. Then when the time lapse completes, all three could arrive at once, this causing an overload for the requesting computer; a situation that leads to congestion and packet loss.
Again according to Cisco, ‘acceptable’ jitter tolerance is as follows:
- Jitter should be below 30 ms.
- Packet loss shouldn’t be more than 1%.
- Network latency should not go over 150ms, (RTT would be 300ms.)
Bandwidth is the maximum rate of data transfer across a given path. The analogy often used refers to the number of lanes in a road – the more lanes the faster the bandwidth
56 kbit/s Modem / Dialup
1.5 Mbit/s ADSL Lite
1.544 Mbit/s T1/DS1
2.048 Mbit/s E1 / E-carrier
4 Mbit/s ADSL1
10 Mbit/s Ethernet
11 Mbit/s Wireless 802.11b
24 Mbit/s ADSL2+
44.736 Mbit/s T3/DS3
54 Mbit/s Wireless 802.11g
100 Mbit/s Fast Ethernet
155 Mbit/s OC3
600 Mbit/s Wireless 802.11n
622 Mbit/s OC12
1 Gbit/s Gigabit Ethernet
1.3 Gbit/s Wireless 802.11ac
2.5 Gbit/s OC48
5 Gbit/s USB 3.0
7 Gbit/s Wireless 802.11ad
9.6 Gbit/s OC192
10 Gbit/s 10 Gigabit Ethernet, USB 3.1
40 Gbit/s Thunderbolt 3
100 Gbit/s 100 Gigabit Ethernet
When all of the above is combined the amount of data that can be sent and received can be calculated. This gives a real world speed of a network connection.
Problem Solving Experience
Everything IT’s top of the range portable test equipment is precisely calibrated to give accurate readings for all of the above, plus, with dedicated modules for network and VoIP specifically designed to interface into ‘live data’ without affecting performance, we routinely solve the most complex of network problems and can install the most reliably functional networks with great ease. Do you need to understand all this? Not at all…but isn’t it a comfort to know that we do!