Ubiquiti LTU | part 2 | professional field tests
- A word of introduction
- Packet efficiency above all else!
- Ubiquiti LTU - long distance test
- 2km test results
- 5km test results
- Ubiquiti LTU - Results 30MHz 2.1-BETA4
- Ubiquiti LTU - Results 30MHz 2.1-BETA5
- Ubiquiti LTU - 50MHz 2.1-BETA4 results
- Ubiquiti LTU - 50MHz 2.1-BETA5 results
- Version 2.1.0-BETA5
- Conclusions
- Summary
A word of introduction
This new solution from Ubiquiti Inc. is gaining more and more recognition among followers of wireless networks, and LTU technology is becoming more and more popular. In previous publications we have already had a chance to familiarize ourselves with the whole range of LTU products and performed preliminary tests of this solution (desktop). This time we will present you with the results of professional tests performed in the field. On this basis, we will draw conclusions on how efficient (packet performance) the devices from Ubiquiti Inc. based on the LTU protocol are. All the obtained results will be available in the further part of this article. We invite you to read it.
Packet efficiency above all else!
Just as a reminder. The main idea of developing the LTU protocol (LTU technology) was to increase network efficiency for PTMP connections! The devices are to work perfectly in harsh environments, where noise is very high. Ubiquiti Inc. developed a new LTU protocol to connect a very large number of clients, while maintaining very high wireless network performance. All of this is possible thanks to several factors:
- 21.2bps per Hz (spectral efficiency)
- over 2 million packets per second
- 4096QAM modulation
Ubiquiti LTU - long distance test
Right. Enough theory! Let's move on to the performance offered by LTU technology. We performed tests for distances above 4.5km (LTU Pro) and below 2km (LTU Lite)
Stress test assumptions for LTU:
Stress test assumptions for LTU:
- CPE end in a LoS range (client cpe) of over 4.5km with a minimum modulation of 64QAM
- Minimum CPE performance with 20MHz channel width at 64QAM Dowload 32Mbps/Upload 7Mbps
- Sector capacity at a 20MHz channel width with 120Mbps/24Mbps minimum
- Measurement of CPE and "sector" performance according to EtherSAM standards (asymmetric traffic) for traffic with IMIX characteristics (Internet MIX according to IETF RFC 6985 "IMIX Genome: Specification of Variable Packet Sizes for Additional Testing") according to parameters presented in the table below.
- The distribution of channels on sectors (4x90 degrees) with no spacing in frequencies. That is, sectors A and C at a center frequency of 5700MHz, and sectors B and D at a center frequency of 5680MHz.
- Tests were performed in the L3 layer
- Packet loss at the level of 0,03%.
- Maximum packet time 50ms
- Maximum packet jitter 40ms
- Maximum instantaneous decrease of preset (minimum) speed to 95%.
- EtherSAM test duration 2 hours
Hardware used for testing
- Ubiquiti LTU Rocket
- Ubiquiti LTU Pro (for tests > 4.5km)
- Ubiquiti LTU Lite (for tests < 2km)
- RF Elements HG3-TP-A90antenna
- VeeEx VePAL TX300S meter with software version 3.5.13 for platform and 3.5.7 for module
The structure of packets used in the test
|
No. |
Bundle length [B]. |
Number of packets (ratio) |
Number of packets for traffic (D/U) 32Mbps/7Mbps |
Number of packets for traffic (D/U) 120Mbps/24Mbps |
|
|
1 |
68 |
7 |
D: 11507 ppsU |
D: 43151 ppsU |
|
|
2 |
595 |
4 |
|||
|
3 |
1518 |
1 |
|||
|
Total one set of packets |
2181 bytes |
12 packages |
2km test results
Now that we know the specifics of our tests, let's move on to the results obtained. First, we will take a closer look at the results we obtained during the 2-hour test with the EtherSAM meter.
Test results for one CPE (LTU Pro) at a distance of about 2km and a frequency of 6100MHz. The height of the CPE is about 5m above sea level, LoS.
Test results for one CPE (LTU Pro) at a distance of about 2km and a frequency of 6100MHz. The height of the CPE is about 5m above sea level, LoS.
We performed further tests with the VeEx VePal TX300S meter at a set throughput of Download 130Mbps, Upload 28Mbps with the IMIX* characteristics described in the table above.
"Local Results" -> uploading from CPE
"Remote Results" -> downloading to CPE.
Network traffic characteristics used for measuring network device performance described in RFC6985 as Internet Mix (IMIX) or other characteristics specified in the manufacturer's documentation.
"Local Results" -> uploading from CPE
"Remote Results" -> downloading to CPE.
Network traffic characteristics used for measuring network device performance described in RFC6985 as Internet Mix (IMIX) or other characteristics specified in the manufacturer's documentation.
5km test results
After successful testing at distances of under 2km, we decided to increase the distance to almost 5km and changed the CPE client device to the Ubiquiti LTU-Pro. What's more! We performed tests at longer distances using different channel widths (20/30/40/50MHz) and different software versions: 2.1-BETA4 and 2.1-BETA5.
Distance and Fresnel Zone
Ubiquiti LTU throughput test results
As I mentioned earlier, we ran the tests for different software versions, different channel widths, and with a variable"Performance Boost" parameter. Below, the results from the obtained Download and Upload speeds.|
Distance |
20MHz |
30MHz |
40MHz |
50MHz |
||||||||||||
|
2.1-BETA4 |
2.1-BETA5 |
2.1-BETA4 |
2.1-BETA5 |
2.1-BETA4 |
2.1-BETA5 |
2.1-BETA4 |
2.1-BETA5 |
|||||||||
|
Performance Boost |
OFF |
ON |
OFF |
ON |
OFF |
ON |
OFF |
ON |
OFF |
ON |
OFF |
ON |
OFF |
ON |
OFF |
ON |
|
Download [Mbps] |
134.7 |
137.8 |
137.2 |
137.9 |
183.0 |
204.8 |
183.5 |
205.4 |
187.1 |
268.5 |
181.8 |
270.2 |
211.3 |
246.8 |
188.3 |
248.5 |
|
Upload [Mbps] |
25.2 |
26.3 |
26.4 |
25.7 |
42.0 |
41.8 |
42.0 |
40.9 |
55.7 |
54.7 |
55.8 |
54.6 |
48.1 |
47.3 |
47.6 |
46.5 |
|
Duplex |
133.9 |
135.2 |
136.5 |
135.3 |
183.8 |
203.1 |
180.5 |
204.2 |
186.6 |
267.7 |
179.4 |
264.9 |
209.1 |
228.7 |
198.8 |
272.9 |
Based on the above results, the conclusions are almost self-evident. The results for 50 MHz are surprisingly low, compared to those for 40 MHz! This is due to lower modulations that the devices have negotiated for this channel width, so it is not recommended to set large channel widths for distant links, unless, of course, the device works as a radio link.
The most optimal is to use 30MHz and we strongly recommend using this channel width. It's definitely a compromise between the effective range and the total capacity of a given sector (transmitter).
Below we present the obtained results. Let's focus only on the tests in the 30 and 50MHz channel widths, since we care about them the most.
Ubiquiti LTU - Results 30MHz 2.1-BETA4
.
Ubiquiti LTU - Results 30MHz 2.1-BETA5
Ubiquiti LTU - 50MHz 2.1-BETA4 results
Ubiquiti LTU - 50MHz 2.1-BETA5 results
Version 2.1.0-BETA5
As can be seen in the above graphics, software version 2.1.-BETA5 does not bring groundbreaking changes in both parameters and throughput. On the other hand, it contains significant improvements in the stability of connections and completely changes the appearance of the graphical interface, so that the GUI works much faster and smoother. Below we present some screenshots of the new Ubiquiti LTU GUI.
Conclusions
1. The software up to version 2.0.6 (inclusive) and BETA versions up to 2.1-BETA3 (inclusive) caused very high stability, performance and interference problems.
2. Enabling the "Performance Boost" function results in increased throughput (maximum band saturation values), at the expense of the number of CPEs the base station can support.
3. Until software versions 2.0.7 and 2.1-BETA4 (inclusive) the split frequency function did not cause any increase in throughput
4. Any client CPE that is not used for a longer period of time (sometimes after only 2 minutes) goes to sleep. In the case of download/upload on software up to 2.0.7 and 2.1-BETA4 (inclusive), with full available bandwidth, this caused a stair-stepping effect and of course also involved an increase in modulation. This effect caused the dormant CPE to lose packets at the very beginning (due to buffering and error correction, this is not visible without advanced meters). Software 2.1-BETA5 significantly improves this unfavorable effect and then the problem is no longer troublesome. Tests with the CPE "awake" showed no packet loss.
2. Enabling the "Performance Boost" function results in increased throughput (maximum band saturation values), at the expense of the number of CPEs the base station can support.
3. Until software versions 2.0.7 and 2.1-BETA4 (inclusive) the split frequency function did not cause any increase in throughput
4. Any client CPE that is not used for a longer period of time (sometimes after only 2 minutes) goes to sleep. In the case of download/upload on software up to 2.0.7 and 2.1-BETA4 (inclusive), with full available bandwidth, this caused a stair-stepping effect and of course also involved an increase in modulation. This effect caused the dormant CPE to lose packets at the very beginning (due to buffering and error correction, this is not visible without advanced meters). Software 2.1-BETA5 significantly improves this unfavorable effect and then the problem is no longer troublesome. Tests with the CPE "awake" showed no packet loss.
5. all Ubiquiti LTU devices, after exceeding the operating frequency of 6000MHz, experience a gradual deterioration of wireless parameters. The difference between the performance of devices operating at 5500MHz and 6200MHz can reach up to 10dBm! Additionally, all CPEs, on the second receiving channel, are noticing increasing difference in comparison to the first receiving channel. As the frequency increases, the parameters of the base station reception deteriorate, which can be seen on the screens below (distance of 4.88km). We have tested it on various combinations of antennas, including Ubiquiti AM-5AC22-45 and Ubiquiti AP-5AC-90-HD.
6. the 50MHz width at 4.88km is too wide to obtain the same modulations as with the narrower channels.
Summary
As a quick recap. When using Ubiquiti LTU solutions you need to pay special attention to the fact that the GPS antenna is sensitive to interference from metal structures. To properly communicate with a large number of satellites, it needs good placement, otherwise we will have problems with gps synchronization. It should also be remembered that an unused CPE "goes to sleep", which limits the range of packet exchange for maintaining the connection. Sudden activation of transmission causes that the device needs "time" to negotiate appropriate bandwidth for itself on the sector. This results in the above mentioned "stairs" at the very beginning of the speed graphs.
One thing is certain. The Ubiquiti LTU series is distinguished by very high packet throughput! And the RoadMap of new features itself makes us optimistic that there will be further changes soon (in future versions of the software), which will increase throughput and operation in nLoS installations. Add to that strong encryption of wireless transmission and the "Performance Boost" mode and we get a very efficient solution that will work in all conditions. And above all, Ubiquiti LTU network devices provide customers with a stable service from the top shelf.
Authors:
Wojciech Repinski
Leszek Błaszczyk
One thing is certain. The Ubiquiti LTU series is distinguished by very high packet throughput! And the RoadMap of new features itself makes us optimistic that there will be further changes soon (in future versions of the software), which will increase throughput and operation in nLoS installations. Add to that strong encryption of wireless transmission and the "Performance Boost" mode and we get a very efficient solution that will work in all conditions. And above all, Ubiquiti LTU network devices provide customers with a stable service from the top shelf.
Authors:
Wojciech Repinski
Leszek Błaszczyk