Emerging Metro Ethernets create new opportunities to converge data and telephony services. Howeve... more Emerging Metro Ethernets create new opportunities to converge data and telephony services. However, to connect legacy telephony equipment, networks have to meet customary QoS requirements even at the presence of bursty data cross traffic. We show with active probing measurements that without appropriate mechanisms the QoS requirements cannot be met. Thus, mechanisms for priorization are needed.
In this thesis, we investigate the problem of channel allocation in a static multihop wireless me... more In this thesis, we investigate the problem of channel allocation in a static multihop wireless mesh network employed for telephony in public safety scenarios. We assume that each mesh router is equipped with two IEEE 802.16-2004 radio interfaces. These interfaces can be switched to one of 12 orthogonal channels. The objective is to find a set of channel allocations for the interfaces in the network that minimize the network interference and fairly distribute capacity. We evaluate three classes of algorithms for channel allocation: Random allocation, greedy allocation and merge allocation with numeric calculations on graphs as well as with simulations in QualNet. We study the following scenarios for distribution of mesh routers: Random placement of nodes with four different numbers of nodes and grid placements of nodes also with four different number of nodes. The numeric calculations approach reveals that all greedy algorithms perform well and limitations in capacity are mainly due to a suboptimal topology construction. The QualNet simulation shows that networks build of routers with IEEE 802.16-2004 interfaces may not have sufficient capacity for the studied public safety scenarios.
Self-organizing wireless metropolitan area networks connect-ing portable GSM/UMTS base stations a... more Self-organizing wireless metropolitan area networks connect-ing portable GSM/UMTS base stations are a valuable asset to rescue forces in disaster recovery. These networks can be formed by routers em-ploying multiple commercially available IEEE 802.16-2004 interfaces. In this paper, we present and evaluate an algorithm for self-organized topol-ogy construction and channel allocation. The algorithm suggests (i) to estimate the capacity of potential wireless links from the measured signal-to-noise ratio, (ii) to run a modified Dijkstra algorithm to select links for topology construction in a way that maximizes network capacity and (iii) to greedily allocate channels to minimize interference. We evaluate our algorithm by simulations using QualNet. With the settings preferred by our industry partner, we find that 12 routers per square kilometer are sufficient to enable coverage. Trading path qualities and traffic load on gateways to the wired network significantly improves performance.
In metropolitan areas the deployment of optical fiber and Gigabit Ethernets leads to an expansion... more In metropolitan areas the deployment of optical fiber and Gigabit Ethernets leads to an expansion of packetswitched networks with large available capacities. These capacities could thus be employed to carry traffic from existing GSM/UMTS base stations and PBX (Private Branch Exchanges). Carrying this traffic requires an emulation of E1/T1 telephone circuits over Ethernets, i.e. E1/T1 PDH/SDH signals from base stations need to be packetized and encapsulated in a circuit emulation adapter (CEA) before being sent over the Ethernet. The key problem with this emulation is to adjust the buffer play-out rate in the receiving CEA to the rate of the sending CEA. This synchronization is necessary: (i) to avoid long-time receiver buffer under-or overflows since base stations are up for weeks and months and (ii) to preserve the frequency of PDH signals across the Ethernet. TIK/ETH and Siemens have started the CoP (Circuit-over-Packets) project to address this problem and to build a new CEA demonstrator. This thesis documents laboratory measurements of the synchronization between pairs of CEA demonstrators over metropolitan Gigabit Ethernets, i.e. we have measured the Maximum Relative Time Interval Error (MRTIE) between the PDH signal that goes into the sending CEA and the PDH signal that comes out of the receiving CEA. To make measurements reproducible, we have employed the network emulator RplTrc [2] to emulate traffic patterns of a Gigabit Ethernets. RplTrc has also been developed by TIK/ETH. To get a feel for the achievements of the CoP project, we have additionally compared measurements of the CoP's CEA to measurements with three commercially available CEAs ("product A, B, C"). From our measurements, we conclude that the CoP's CEA generally outperforms the CEA C. The CoP's CEA also shows better performance in long-term synchronization stability than the CEA B. However, the synchronization algorithm of the CEA B is more robust to variable network conditions. Finally, we have found that the CEA A shows higher quality level for short-term synchronization stability than the CoP's CEA, while long-term behaviors are comparable. The CEA A is slightly more robust to different traffic scenarios, but equally sensitive to the limited delay (im)precision of RplTrc [2]. We explain this slightly better performance of the CEA A with the arithmetic limitations of the Siemens board used to implement the CoP's CEA demonstrator.
The question of our study is how to provision a diffserv (differentiated service) intra-net servi... more The question of our study is how to provision a diffserv (differentiated service) intra-net serving three classes of traffic, i.e., voice, real-time data (e.g. stock quotes), and best-effort data. Each class of traffic requires a different level of QoS (Quality of Service) guarantee. For VoIP the primary QoS requirements are delay and loss; for realtime data response-time. Given a network configuration and anticipated workload of a business intra-net, we use ns-2 simulations to determine the minimum capacity requirements that dominate total cost of the intranet. To ensure that it is worthwhile converging different traffic classes or deploying diffserv, we cautiously examine capacity requirements in three sets of experiments: three traffic classes in i) three dedicated networks, ii) one network without diffserv support , and iii) one network with diffserv support. We find that for the business intra-net of our study, integration without diffserv may need considerable over-provisioning depending on the fraction of real-time data in the network. In addition, we observe significant capacity savings in the diffserv case; thus conclude that deploying diffserv is advantageous. The relations we find give rise to, as far as we know, the first rule of thumb on provisioning a diffserv network for increasing real-time data.
Proceedings of the 3rd international conference on Quality of service in heterogeneous wired/wireless networks - QShine '06, 2006
creates new opportunities to save costs by converging data and telephone services. The primary qu... more creates new opportunities to save costs by converging data and telephone services. The primary question of our study is whether emerging metropolitan networks can meet the QoS requirements necessary to connect GSM and UMTS base stations. These requirements on delay, jitter, and loss are significantly more stringent than the ones for VoIP and have to be met at the presence of bursty cross traffic. Therefore, we have probed ETH's campus network which spans the metropolitan area of Zurich with traffic comparable to encapsulated E1 traffic from base stations and have measured perceived QoS. Our findings show that the campus network generally meets the QoS requirements. However, the perceived QoS degrades with increasing network utilization. To further investigate the impact of the configuration of Ethernet switches on perceived QoS, we have conducted a simulation study. The results show that the perceived QoS is better when switch buffers are limited to small sizes of 1 MB compared to setups where the number of frames in the buffer is limited. From these results we infer that metropolitan Gigabit Ethernets are well suited for connecting GSM and UMTS base stations.
Emerging Metro Ethernets create new opportunities to converge data and telephony services. Howeve... more Emerging Metro Ethernets create new opportunities to converge data and telephony services. However, to connect legacy telephony equipment, networks have to meet customary QoS requirements even at the presence of bursty data cross traffic. We show with analytical evaluation and simulation results that without appropriate mechanisms the QoS requirements cannot be met.
Proceedings of the 2006 SIGCOMM workshop on Large-scale attack defense, 2006
The security level of networks and systems is determined by the software vulnerabilities of its e... more The security level of networks and systems is determined by the software vulnerabilities of its elements. Defending against large scale attacks requires a quantitative understanding of the vulnerability lifecycle. Specifically, one has to understand how exploitation and remediation of vulnerabilities, as well as the distribution of information thereof is handled by industry. In this paper, we examine how vulnerabilities are handled in large-scale, analyzing more than 80,000 security advisories published since 1995. Based on this information, we quantify the performance of the security industry as a whole. We discover trends and discuss their implications. We quantify the gap between exploit and patch availability and provide an analytical representation of our data which lays the foundation for further analysis and risk management.
Evaluating the performance of network sensitive applications, devices, and protocols has become i... more Evaluating the performance of network sensitive applications, devices, and protocols has become increasingly complex,as the diversity of scenarios and speed of networks increase. This includes the evaluation of VoIP and video conferencing applications, telephony devices, and protocols in both wired and wireless scenarios. Therefore, in this paper, we describe how we design and implement ReplTrc, a network emulation tool that is capable to replay large packet delay traces with high accuracy in timing to enable performance,evaluation. We assume that packet traces have previously been generated or captured with network probing, simulation, or network calculus. Thus, evaluations with ReplTrc account for all important performance,characteristics such as long-range dependence and self-similarity of traffic or cross-traffic that are refle cted in the packet traces. ReplTrc essentially consists of (i) a kernel module which intercepts the Linux protocol stack to delay, drop, or duplicate pac...
Packet switching technologies such as Gigabit Ethernet or IP over WDM are gradually replacing the... more Packet switching technologies such as Gigabit Ethernet or IP over WDM are gradually replacing the traditional telephony (PDH/SDH) networks due to their ability to transport data traffic at very high rates. Connecting UMTS/GSM base stations over Metropolitan Gigabit Ethernets has a great potential of cost savings. In our scenario telephony base stations are connected to the core PDH network through a Metropolitan Gigabit Ethernet. This requires an emulation of E1/T1 telephone circuits over Gigabit Ethernets, i.e. E1/T1 PDH/SDH signals from base stations need to be packetized and encapsulated in a circuit emulation adapter (CEA) before being send over the Ethernet. After receiving the signals have to be reassembled. In contrast to VoIP, these emulated connections stay up for weeks. Thus, the on going frequency synchronization between the sender and the receiver clock is required. Much research has been conducted on the problem of clock synchronization. However, known algorithms such as the NTP and the Linear Regression do not meet stringent requirements of ITU-T G.823. In this report we present and evaluate open source implementations of synchronization algorithms: Convex Hull and Linear Adaptive Approximation. These algorithms are based on the concept of estimating the minimal end-to-end delay from sender and receiver time stamps. We show that such implementations meet the ITU-T G.823 requirements for synchronization and generally outperform the synchronization algorithms employed in commercially available circuit emulation adapters.
The test environment for sensitive network applications needs to be flex-ible. This is only possi... more The test environment for sensitive network applications needs to be flex-ible. This is only possible in a lab setting containing a suitable network emulator. This thesis describes different possibilities for handling network emulation focussing mainly on emulators running on commodity ...
Uploads
Papers by Ulrich Fiedler