[Knowledge Sharing] LTE Latency Improvement Gains

Packet data latency is a key performance metric in today’s communication systems that is regularly measured by vendors and operators but also end-users, e.g. via speed test applications.

Latency measurements are done in all phases of the lifetime of a radio access network system; starting when verifying a new software release or system component, and continuing when deploying a system and after the system is put in commercial operation.

Figure 1. Example presentation of typical speed test application.

LTE was designed with low latency in mind already from the beginning, and as a result today LTE does indeed have better packet data latency than previous generations of the 3GPP RATs. Also, by a wide range of end-users LTE is now recognized to be a system that provides faster access to internet and lower data latencies than previous generations of mobile radio technologies.

Since the introduction of LTE in 2009, several improvements had been developed, however, mainly targeting the increase of the maximum data rates of the system, e.g. Carrier Aggregation, 8×8 MIMO, etc. To also get full benefit of these data rate enhancements; we strongly believe that continuous enhancements of the latency of LTE should also be an important part of the future evolution track of LTE.

Packet data latency is a parameter that indirectly influences the perceived data rate of the system. HTTP/TCP is the dominating application and transport layer protocol suite used on the internet today. According to HTTP Archive (http://httparchive.org/trends.php), the typical size of HTTP based transactions over the internet are in the range of a few 10’s of Kbytes up to 1 Mbyte. In this size range, the TCP slow start period is a significant part of the total transport period of the packet stream.

During TCP slow start, TCP exponentially increases its congestion window, i.e. the number of segments it brings into flight, until it fully utilizes the throughput LTE can offer. The incremental increases are based on TCP ACKs which are received after one round trip delay in the LTE system. Thus, as it turns out, and is shown below, during TCP slow start the performance is latency limited also in LTE. Hence, improved latency in LTE can improve the perceived data rate for TCP based data transactions, which in its turn reduces the time it takes to complete a data down- or upload.

Figure 2. The initial TCP slow start phase; the ramp up of number of bytes sent (=number of TCP segments sent) is depending on the round trip delay until the transmitter gets the acknowledgments (TCP ACKs) for the transmitted bytes.

So what areas can be addressed in LTE to reduce packet latency and what are the gains when doing so? In this blog post we will have a look at three different improvement areas, which are also further illustrated in Figure 3.

The main areas are:

the uplink access procedure that starts with the need for a UE to request the permission of the base station to send (i.e. the Scheduling Request (SR) procedure),
the standardized processing times in LTE that stipulate that for instance retransmissions of data will (at least) take 8 ms
the current transmission time interval (TTI) of 1 ms

Figure3. Example of areas that can be addressed for LTE latency reductions.

Figure 4 below shows the potential gains in terms of increased download throughput and reduction of download time for the latency reduction techniques of allowing for instant access in the uplink optimizing the latency of the SR procedure, and TTI shortening (to ½) including shortening of the processing delays to half, as well as a combined approach. In the simulation we analyzed download file sizes (e.g. of an HTTP item) of 100 KB and 1MB. With instant access in the uplink (= allowing the UE to transmit when it gets data on a pre-established grant) or by reducing the TTI and processing times to 1/2ms the download performance can be improved significantly, i.e. the maximum LTE link throughput of 150Mbit/s as assumed in this example can be utilized more effectively

Figure 4. Simulated download speed (x-axis in Mbit/s) and relative reduction of download time (as %) for different file sizes and latency reduction techniques. The maximum achievable LTE link throughput is assumed to be 150Mbit/s in this simulation and the transport network delay is omitted.

The radio roundtrip reductions when introducing the possibility of instant access in the uplink is 8 ms or more (i.e. the SR-to-grant roundtrip delay is removed, which is at least 8 ms) , and when reducing TTIs and processing time by half, the reduction is about 4 ms. But despite these rather small (in terms of milliseconds) improvements of the radio round trip time, the total increase in in the perceived throughput and delay savings of downloading an item below 1MB is significant due to the additive effect of LTE latency improvements in the TCP slow start. It should, however, be noted that the gain is a function of the transport network delay in the system, and here we show the best possible gains when omitting the transport network latency.

Latency improvements may not only turn out to be beneficial for TCP down- and uploads. Another example of a benefit could be that radio resource efficiency could be positively impacted by latency reductions. Lower packet data latency could increase the number of transmissions possible within a certain delay bound; hence higher-rate transmissions (higher MCS due to higher BLER targets) could be used for the data transmissions thereby freeing up radio resources and potentially improving the capacity of the system.

Furthermore, there are a number of current applications that will even directly benefit from reduced latency in terms of increased perceived quality of experience: examples are gaming, real-time applications like VoLTE/OTT VoIP, and video conferencing.

And going even beyond the current use cases, into the future of the networked society, there will be a number of new applications that will benefit from or even require reduced latencies. Examples may be remote control/driving of vehicles, augmented reality applications in e.g. smart glasses, or specific machine type communications.

We hope you found this blog post interesting and that it inspired you to think about latency reduction techniques for LTE, or what benefits a lower packet data latency in LTE would bring to your favorite application.

Per Synnergren Principal Researcher, Ericsson Research
Torsten Dudda Experienced Researcher, Ericsson Research

Source : https://www.ericsson.com/research-blog/lte/lte-latency-improvement-gains/