Brandon Heller

A network in a laptop: rapid prototyping for software-defined networks

Mininet is a system for rapidly prototyping large networks on the constrained resources of a single laptop. The lightweight approach of using OS-level virtualization features, including processes and network namespaces, allows it to scale to hundreds of nodes. Experiences with our initial implementation suggest that the ability to run, poke, and debug in real time represents a qualitative change in workflow. We share supporting case studies culled from over 100 users, at 18 institutions, who have developed Software-Defined Networks (SDN). Ultimately, we think the greatest value of Mininet will be supporting collaborative network research, by enabling self-contained SDN prototypes which anyone with a PC can download, run, evaluate, explore, tweak, and build upon.

The controller placement problem

Network architectures such as Software-Defined Networks (SDNs) move the control logic off packet processing devices and onto external controllers. These network architectures with decoupled control planes open many unanswered questions regarding reliability, scalability, and performance when compared to more traditional purely distributed systems. This paper opens the investigation by focusing on two specific questions: given a topology, how many controllers are needed, and where should they go? To answer these questions, we examine fundamental limits to control plane propagation latency on an upcoming Internet2 production deployment, then expand our scope to over 100 publicly available WAN topologies. As expected, the answers depend on the topology. More surprisingly, one controller location is often sufficient to meet existing reaction-time requirements (though certainly not fault tolerance requirements).

Reproducible network experiments using container-based emulation

In an ideal world, all research papers would be runnable: simply click to replicate all results, using the same setup as the authors. One approach to enable runnable network systems papers is Container-Based Emulation (CBE), where an environment of virtual hosts, switches, and links runs on a modern multicore server, using real application and kernel code with software-emulated network elements. CBE combines many of the best features of software simulators and hardware testbeds, but its performance fidelity is unproven. In this paper, we put CBE to the test, using our prototype, Mininet-HiFi, to reproduce key results from published network experiments such as DCTCP, Hedera, and router buffer sizing. We report lessons learned from a graduate networking class at Stanford, where 37 students used our platform to replicate 18 published results of their own choosing. Our experiences suggest that CBE makes research results easier to reproduce and build upon.

Where is the debugger for my software-defined network?

The behavior of a Software-Defined Network is controlled by programs, which like all software, will have bugs - but this programmatic control also enables new ways to debug networks. This paper introduces ndb, a prototype network debugger inspired by gdb, which implements two primitives useful for debugging an SDN: breakpoints and packet backtraces. We show how ndb modifies forwarding state and logs packet digests to rebuild the sequence of events leading to an errant packet, providing SDN programmers and operators with a valuable tool for tracking down the root cause of a bug.

Divided Edge Bundling for Directional Network Data

The node-link diagram is an intuitive and venerable way to depict a graph. To reduce clutter and improve the readability of node-link views, Holten & van Wijks force-directed edge bundling employs a physical simulation to spatially group graph edges. While both useful and aesthetic, this technique has shortcomings: it bundles spatially proximal edges regardless of direction, weight, or graph connectivity. As a result, high-level directional edge patterns are obscured. We present divided edge bundling to tackle these shortcomings. By modifying the forces in the physical simulation, directional lanes appear as an emergent property of edge direction. By considering graph topology, we only bundle edges related by graph structure. Finally, we aggregate edge weights in bundles to enable more accurate visualization of total bundle weights. We compare visualizations created using our technique to standard force-directed edge bundling, matrix diagrams, and clustered graphs; we find that divided edge bundling leads to visualizations that are easier to interpret and reveal both familiar and previously obscured patterns.

Optimizing a virtualized data center

Many data centers extensively use virtual machines (VMs), which provide the flexibility to move workload among physical servers. VMs can be placed to maximize application performance, power efficiency, or even fault tolerance. However, VMs are typically repositioned without considering network topology, congestion, or traffic routes. In this demo, we show a system, Virtue, which enables the comparison of different algorithms for VM placement and network routing at the scale of an entire data center. Our goal is to understand how placement and routing affect overall application performance by varying the types and mix of workloads, network topologies, and compute resources; these parameters will be available for demo attendees to explore.

Using Network Knowledge to Improve Workload Performance in Virtualized Data Centers

The scale and expense of modern data centers motivates running them as efficiently as possible. This paper explores how virtualized data center performance can be improved when network traffic and topology data informs VM placement. Our practical heuristics, tested on network-heavy, scale-out workloads in an 80 server cluster, improve overall performance by up to 70% compared to random placement in a multi-tenant configuration.

Green enterprise computing data: Assumptions and realities

Until now, green computing research has largely relied on few, short-term power measurements to characterize the energy use of enterprise computing. This paper brings new and comprehensive power datasets through Powernet, a hybrid sensor network that monitors the power and utilization of the IT systems in a large academic building. Over more than two years, we have collected power data from 250+ individual computing devices and have monitored a subset of CPU and network loads. This dense, long-term monitoring allows us to extrapolate the data to a detailed breakdown of electricity use across the buildings computing systems. Our datasets provide an opportunity to examine assumptions commonly made in green computing. We show that power variability both between similar devices and over time for a single device can lead to cost or savings estimates that are off by 15-20%. Extending the coverage of measured devices and the duration (to at least one month) significantly reduces errors. Lastly, our experiences with collecting data and the subsequent analysis lead to a better understanding of how one should go about power characterization studies. We provide several methodology guidelines for future green computing research.