Jarmo Prokkola

HSDPA Performance in Live Networks

The first HSDPA (high-speed downlink packet access) networks have been recently deployed for operational use. We evaluate and compare live HSDPA operational network performance from the end-user perspective, especially related to voice over IP and Web browsing applications. We compare goodput performance with user datagram protocol (UDP) and transmission control protocol (TCP) in both wideband code division multiple access (WCDMA) and HSDPA. We also address one-way delay measurements to show how up- and downlink asymmetry affects delay and jitter. We found that HSDPA and WCDMA goodput correlate well with the advertised maximum performance values. However, with small payload sizes HSDPA does not bring a substantial improvement over TCP. With UDP the situation is clearly better except with payload sizes commonly used by VoIP. We verify that HSDPA delay is substantially smaller than with WCDMA, as well as the occurrence of delay spikes caused by link level retransmissions. We found that despite 2 ms time division in HSDPA, the main jitter level is at 10 ms. The results can be exploited by operators and equipment manufacturers for service optimisation.

Measuring WCDMA and HSDPA Delay Characteristics with QoSMeT

Quality of service (QoS) is becoming increasingly important with the rise of multimedia applications (e.g., voice over IP (VoIP), video conferencing, online gaming, and Internet protocol television (IPTV)) in public data networks. These applications demand real-time service, while the networks should still be able to transfer also other traffic reliably (e.g., data, e- mail, WWW). We have developed a measurement tool, which is able to measure one-way QoS performance statistics. The tool enables passive monitoring of the desired application(s) and gives information about the QoS provided by the network. The tool can be used practically over any kind of network structure as long as IP is supported. In this paper, we focus on detailed end-to-end delay characteristics measured in live HSDPA (high speed downlink packet access) enabled 3G network. HSDPA provides clearly better delay performance than the basic WCDMA (wideband code division multiple access). Interestingly, deterministic long time behavior in delay was observed, and also, discrete jitter levels were found. With argumentation, and finally with a trial, we proved that these phenomena originate from the CBR-type (constant bit rate) VoIP source traffic affecting with 3G network timing mechanisms and mutual clock drifts. As a side product, the study also showed that it is advisable to verify the operation of traffic generation tools before using them.

Cognitive Radio Trial Environment: First Live Authorized Shared Access-Based Spectrum-Sharing Demonstration

Cognitive radio system (CRS) technology can help respond to the growing mobile traffic demand by improving network resource usage and gaining access to new shared spectrum bands. This article presents a cognitive radio trial environment (CORE) consisting of cognitive engines (CEs) to control different radio systems [e.g., long-term evolution (LTE) and wireless open-access research platform (WARP)-based networks]. Load balancing and authorized shared access (ASA) are demonstrated using the trial environment with promising results. The ASA-based spectrum sharing trial is presented for the first time with a real-life mobile network accessing a shared spectrum band under a licensed shared access (LSA) regime. Cognitive decision making brings intelligence to the usage of the radio and network resources and, at best, increases considerably end users? quality of service (QoS) compared to the standard systems, as shown for QoS-based offloading.

3G/HSPA Performance in Live Networks from the End User Perspective

Operators around the world are improving their 3G/UMTS networks by introducing HSPA (High Speed Packet Access) that includes both, enhanced uplink and downlink. The HSPA is expected to finally provide the mobile broadband access that is able to compete with the fixed connections in performance regarding popular applications such as web browsing, VoIP, and video. However, it has remained unclear how well the live networks fulfill the promises of performance. We contribute in filling this void by providing measurements in live 3G/HSPA networks. We compare TCP and UDP goodput performance in basic WCDMA, HSDPA-only, and HSPA. Moreover, one-way delay and jitter measurement results are presented in a stationary as well as in a mobile scenario. The results show that the enhanced network outperforms clearly its predecessor by offering considerably higher data rates, lower delay, and lower jitter. Also, because of the HARQ (Hybrid Automatic Repeat request), there are notably less delay spikes observed with HSPA than with WCDMA. However, the drive tests show that handovers result in high jitter and interruptions to the communications, which causes, e.g., decreased VoIP call quality. The uplink enhancement (HSUPA) improves the TCP performance beyond the HSDPA-only access, but it is still behind that of the fixed connections. In addition, the properties of the channel allocation mechanism still considerably deteriorate the data rates seen by the user. Nevertheless, the HSPA is an enabler for true mobile broadband internet access.

Performance of IEEE 802.11 based WLAN devices under various jamming signals

IEEE 802.11 based WLAN systems have gained interest to be used in the military and public authority environments, where the radio conditions can be harsh due to intentional jamming. The radio environment can be difficult also in commercial and civilian deployments since the unlicensed frequency bands are crowded. To study these problems, we built a test bed with a controlled signal path to measure the effects of different interfering signals to WLAN communications. We use continuous wideband noise jamming as the point of comparison, and focus on studying the effect of pulsed jamming and frequency sweep jamming. In addition, we consider also medium access control (MAC) interference. Based on the results, WLAN systems do not seem to be sensitive to the tested short noise jamming pulses. Under longer pulses, the effects are seen, and long data frames are more vulnerable to jamming than short ones. In fact, even a small amount of long frames in a data stream can ruin the performance of the whole link. Under frequency sweep jamming, slow sweeps with narrowband jamming signals can be quite harmful to WLAN communications. The results of MAC jamming show significant variation in performance between the different devices: The clear channel assessment (CCA) mechanism of some devices can be jammed very easily by using WLAN-like jamming signals. As a side product, the study also revealed some countermeasures against jamming.

A detailed study of a CDMA based approach to enhance ad hoc network performance

Random access schemes operate typically on a contention based common channel, which brings problems with increasing traffic load. These problems are emphasized in a multi-hop wireless ad hoc network environment. Efficient collision avoidance methods are needed, but they also tend to decrease the spatial capacity of the network. With CDMA (Code Division Multiple Access), it is possible to make multiple simultaneous co-located successful transmissions and thus increase network capacity within the limits of multiple access interference (MAI). Bi-code channel access (BCCA) is a method to apply CDMA in ad hoc networking, providing also a common access channel for network connectivity maintenance and self-configuration. With BCCA, collisions are rare, and thus, more straightforward medium access control (MAC) methods can be used. A MAC solution designed especially for BCCA (BC-MAC) and ad hoc networking increases the efficiency of the channel usage. As BCCA is based on receiver code CDMA, the spreading code of the next hop node is needed. A novel network layer spreading code distribution (NSCD) method is proposed for this purpose. The spreading codes are distributed within routing protocol control packets at the route establishment phase. A detailed study on the performance behavior of different methods is provided. Also, interesting aspects considering the ad hoc network operation are discussed. Instead of the widely used unrealistic cut propagation model, a propagation model with realistic MAI calculation is used in this study. It is shown that the use of NSCD increases the network control load only slightly, and that the performance is practically the same as it is with the assumption of known spreading codes. BCCA with its dedicated MAC and NSCD clearly outperforms commonly used methods (like IEEE 802.11) and yet is easy to implement, robust for design parameters under different conditions while maintaining the original idea of ad hoc networking.

A network information service for quality-driven mobility

Fast growing data traffic over mobile networks pose many challenges for both end systems and networks to satisfy the required service levels of different applications. For end-users, the quality of experience (QoE), and consequently the underlying quality of service (QoS) are the most meaningful criteria for triggering handovers and selecting target base stations. However, difficulties in reliably measuring quality real-time have resulted in many QoS/QoE based mobility solutions being impractical. We propose a network information service that allows mobile devices to find suitable nearby base stations by indicating their capability to satisfy the required service level. Information services allow end systems to discover heterogeneous networks at or near their location. However, the current solutions do not enable assessing base station availability and their quality of service in different locations reliably enough. In our solution, end system performed QoS measurement (and possibly QoE estimation) results are stored in a spatial database within the information service as polygon geometries. Through a single query to the information service end systems can find a sufficient amount of information about base stations in range to make handover target decisions on quality of service basis. We observe that the solution can achieve quality improvements of up to 50% in dense heterogeneous networks.

Supporting IP end-to-end QoS architectures at vertical handovers

IP end-to-end QoS mechanisms are quite important for users of real-time multimedia applications. Wireless access systems are widely used also with multimedia applications although the access networks may present heterogeneous technologies and diverse QoS mechanisms. Vertical seamless handover between different wireless systems is possible while the Internets QoS mechanisms are difficult to integrate into the handover procedures. ITEAs Easy Wireless project develops a new approach to end-to-end QoS in wireless heterogeneous networks. In this paper we discuss how the IP end-to-end QoS architectures can be supported when a seamless vertical handover is executed between different network systems. We focus on analyzing the latest standardized versions of the existing wireless 2/sup nd/ and 3/sup rd/ generation cellular systems together with WLAN or WMAN technologies. Differentiated Services and Integrated Services are compared and Differentiated Services is found to be more appropriate to support widely roaming users.

Preliminary big data in a 5G test network

The Finnish 5G test network (5GTN) is built to support development and analysis of 5G technologies and new services on top of them. 5GTN includes a holistic monitoring and test generation architecture, enabling generation and collection of diverse test and datasets, analyzing the results and using these as basis for developing new technologies, algorithms, services, and optimizations. Both the combined network and service parameters space, as well as data volume, provide us with several big data problems. In this paper, we describe the overall system architecture, the types of data we collect, the initial analysis we have performed on the data, and what we have learned so far. The network is constantly evolving with new services integrated and tested, and new big data analytics being applied as we learn from our tests. As such, these are preliminary results for ongoing work.

DockingAssist: A novel vessel navigation system design based on WiMAX and DGNSS

The DockingAssist system is a centralised, cost-effective, real-time navigation system that provides the necessary centimetre positioning and speed accuracy to ensure efficient and safe manoeuvring within a entire port area, thereby enhancing vessel trajectory and providing constant monitoring of moored/docked vessels. By reducing transit, this system will result in improved port traffic management and lower operating expenses, CO2 emissions and fuel usage, thereby lessening the environmental impact of shipping. The reduced transit time will increase throughput in ports with a low investment. Our solution is based on WiMAX technology transmitting DGNSS positioning data in harbours. We provide proof-of-concept results based on empirical measurements performed in a real harbour. We find that WiMAX fulfils the requirements for accurate vessel positioning and navigation in harbour environment while providing wireless network capacity also for other services.