Investigating Latency in gnu software Radio with usrp embedded Series sdr platform

platform. 
Index Terms
—
Software Defined Radio, SDR, GNU Radio, 
USRP, E100, Delay, Latency. 
I.
I
NTRODUCTION
A software-defined radio (SDR) is a radio communication 
technology that implements most of conventional radio 
components, such as detectors, modulation, de-modulation, 
amplifiers, filters and mixers, in software. Therefore, the key 
advantage of a SDR system is the flexibility. It is easy and 
convenient to prototype, design, develop, test, modify and 
configure on the system. Because of these attractive features, 
SDR systems are considered as a prospective solution for some 
specific scenarios where the radio communication channels are 
easily destroyed or damaged. In these cases, the SDR systems 
automatically adapt themself to maintain the necessary 
communications. However, SDR systems have to overcome a 
big challenge on the performance and real-time processing 
before deploying in the real world. 
The software side of a SDR platform is normally based on 
three approaches: programmable hardware such as field 
programmable gate arrays (FPGAs); embedded digital signal 
processors (DSPs); and general-purpose processor (GPP). It is 
also possible to combine of these three architectures. 
SDR platforms based on FPGA and DSPs satisfy real-time 
processing and timing requirements of most of current wireless 
protocols since complicated signal processing tasks are 
implemented in hardware [4, 5]. However, they are not flexible 
enough and also require broad knowledge on embedded 
techniques and hardware architectures, thus, resulting in longer 
development cycle and higher level of efforts. In fact, these 
systems such as Wireless Open Access Research Platform 
(WARP) are not popular and just for the purpose of research. 
Contrarily, SDR platform architectures based on general-
purpose processor (GPP) like personal computers are really 
flexible. However, this architecture comprises of multiple 
heterogeneous processing units with interconnecting buses 
resulting in introducing high jitter and delay as well as limited 
performance. For instance, the SDR platform GNU Software 
Radio running on an Intel Pentium IV GPP connected with 
USRP1 achieves 8Msps maximum throughput. This is because 
of the bus communication bottleneck since the USB 2.0 only 
sustains maximum data rate at 32Mbps [5, 7, 8, 9, 10, 12, 14]. 
Moreover, the delay of this system reaches to 36.9ms that 
cannot satisfy strict timing and performance requirements of 
current 802.11 wireless protocols [8]. 
Recently, Ettus Research introduced new USRP Embedded 
series (USRP E series) with an approach of combining GPP 
and Embedded DSP which integrates the general functionality 
of a host computer to an embedded processor. That features 
allows the USRP to operate in a standalone fashion. Instead of 
communicating with the PC through a communication bus (e.g 
USB 2.0 and Gigabit Ethernet), USRP E series has a General 
Purpose Memory Controller (GPMC) bus controller for 
connecting between the embedded memory and FPGA in the 
USRP [13]. This new USRP architecture does need to be 
investigated to figure out its properties, particularly the latency. 
2013 Eighth International Conference on Broadband, Wireless Computing, Communication and Applications
978-0-7695-5093-0/13 $31.00 © 2013 IEEE
DOI 10.1109/BWCCA.2013.11
9
2013 Eighth International Conference on Broadband, Wireless Computing, Communication and Applications
978-0-7695-5093-0/13 $31.00 © 2013 IEEE
DOI 10.1109/BWCCA.2013.11
9

---

This article is to examine the latency on the SDR platform 
–
GNU Software Radio running on USRP E100 to find out the 
sources of the delay as well as quantify them. 
II.
B
ACKGROUND AND 
R
ELATED 
W
ORK
Some research measured the latency on GNU Radio/USRP 
platforms [7] but no one used USRP E100 as the hardware part 
of the SDR platforms. Moreover, most of them did not go into 
details of analyzing the sources of the delay.