Torsten Runge

Gas chromatographic enantioseparation of allenes

Abstract Enantioselective gas chromatography with modified cyclodextrins as stationary phases is successfully used to separate chiral allenic hydrocarbons - including kinetically labile 1,2-cyclooctadiene - and 1-halo-1,2-butadienes.

A study of networking software induced latency

For long time, high-speed packet processing has been reserved for specialized hardware devices since software based solutions were not able to achieve the required performance. However, off-the-shelf packet processing hardware and software improved over the last years, which is why software based solutions cope with high-speed traffic nowadays. Due to the flexibility of software there is a trend towards doing packet processing in software, e.g. using OpenFlow or virtual switches. Although packet processing in software offers many capabilities, the complexity of such software bases solutions makes it hard to evaluate, optimize, or predict the networking performance of servers, end user hosts, or routers. We present a study that investigates the packet latency caused by the packet processing in the Linux network stack. We develop a simulation model in ns-3 for packet processing via the Linux network stack that helps understanding of its performance implications. We validate our simulation model based on measurements with nanosecond accuracy and software profiling.

Optimizing latency and CPU load in packet processing systems

High-speed network cards supporting 10 or 40GbE (Gigabit Ethernet) are available today. Software frameworks for high-speed packet reception and transmission were created to exhaust the performance of these cards. However, these frameworks are not applicable as general-purpose solution. Thus, it is necessary to revisit general purpose network IO software that was designed more than a decade ago. In standard Linux settings, connectivity between applications and physical networks happens via the New API (NAPI). This motivated us to investigate how underlying NIC drivers can be adapted to improve latency in combination with the Linux NAPI. Based on testbed measurements, we propose an optimized algorithm for the NIC driver to dynamically adapt the Interrupt Throttling Rate (ITR). We implemented the algorithm and evaluated it with latency and throughput measurements based on the Linux module of Open vSwitch that operates on top of the NAPI. Our measurements show that our new ITR algorithm improves the packet latency without affecting the CPU load as much as other solutions.

Low latency network traffic processing with commodity hardware

Packet processing on commodity hardware is a cost-efficient and flexible alternative to specialized networking hardware. In case of Linux, the classical QoS mechanisms (e.g DiffServ) assume that the outgoing link is the bottleneck. However, on commodity hardware the CPU typically becomes the bottleneck in packet processing. Taking into account current trends in the Internet, we assume that the percentage of latency-sensitive applications (e.g. VoIP, video conferencing, online gaming) will increase. Thus, we propose to extend the Linux NAPI with respect to preferring latency-sensitive traffic at the ingress before reaching the CPU. In this paper, we investigate a new concept for the Linux NAPI for low latency packet processing. Based on a model of a Linux software router, we show in a case study that our concept strongly improves the packet latency of real-time packets at acceptable low costs regarding the achievable maximum throughput. Our model is calibrated and validated based on real testbed measurements (e.g. profiling).

Building a Low Latency Linux Software Router

Packet processing (e.g. routing, switching, firewall) with commodity hardware is a cost-efficient and flexible alternative to specialized networking hardware. On commodity hardware the CPU typically becomes the bottleneck in packet processing. However, in well-known QoS mechanisms (e.g. DiffServ), the outgoing link is assumed to be the bottleneck. This limitation is unfavorable, in particular for latency-sensitive applications (e.g. VoIP, video conferencing, online gaming). Thus, we propose and implement a QoS concept for a Linux software router to prioritize latency-sensitive traffic at the incoming network interface. Our testbed measurements show that our prototype implementation improves the packet processing w.r.t the latency of latency-sensitive traffic even under high traffic loads.

Towards Low Latency Software Routers

Network devices based on commodity hardware are capable of high-speed packet processing while maintaining the programmability and extensibility of software. Thus, software-based network devices, like software routers, software-based firewalls, or monitoring systems, constitute a cost-efficient and flexible alternative to expensive, special purpose hardware. The overall packet processing performance in resource-constrained nodes can be strongly increased through parallel processing based on off-theshelf multi-core processors. However, synchronization and coordination of parallel processing may counteract the corresponding network node performance. We describe how multi-core software routers can be optimized for real-time traffic by utilizing the technologies available in commodity hardware. Furthermore, we propose a low latency extension for the Linux NAPI. For the analysis, we use our approach for modeling resource contention in resource-constrained nodes which is also implemented as a resource-management extension module for ns-3. Based on that, we derive a QoSaware software router model which we use to evaluate our performance optimizations. Our case study shows that the different scheduling strategies of a software router have significant influence on the performance of handling realtime traffic.

A Modeling Approach for Resource Management in Resource-Constrained Nodes

The rapid growth of link bandwidths on one hand, and the emergence of resource-constrained nodes (e.g. software routers) on the other hand, will cause network nodes to be the bottleneck in the future. Parallel processing using multi-core processors can increase the packet processing of resource-constrained nodes and alleviate the problem. However, intra-node resource contention can have a strong negative impact on the corresponding network node and, therefore, also on the overall performance of the network. Commonly used network simulators (e.g. ns-3) only offer a rather simplistic node model and do not take into account intra-node resource contention. We propose a unified and extensible approach to model intra-node resource management in resource-constrained nodes. Our model gives ability to identify and predict performance bottlenecks in networks. We have implemented our model as an extension to the network simulator ns-3. The simulation results using different case studies, show that our approach significantly outperforms the original ns-3 in terms of realistic modeling.