Cloud resilience is crucial for cloud operators and the myriad of applications that rely on the cloud. Today, we lack a mechanism that enables cloud operators to perform graceful degradation of applications while satisfying the application's availability requirements. In this paper, we put forward a vision for automated cloud resilience management with cooperative graceful degradation between applications and cloud operators. First, we investigate techniques for graceful degradation and identify an opportunity for cooperative graceful degradation in public clouds. Second, leveraging criticality tags on containers, we propose diagonal scaling---turning off non-critical containers during capacity crunch scenarios---to maximize the availability of critical services. Third, we design Phoenix, an automated cloud resilience management system that maximizes critical service availability of applications while also considering operator objectives, thereby improving the overall resilience of the infrastructure during failures. We experimentally show that the Phoenix controller running atop Kubernetes can improve critical service availability by up to 2× during large-scale failures. Phoenix can handle failures in a cluster of 100,000 nodes within 10 seconds. We also develop AdaptLab, an open-source resilience benchmarking framework that can emulate realistic cloud environments with real-world application dependency graphs.

Internet measurement research faces an accessibility crisis: complex analyses require custom integration of multiple specialized tools that demands specialized domain expertise. When network disruptions occur, operators need rapid diagnostic workflows spanning infrastructure mapping, routing analysis, and dependency modeling. However, developing these workflows requires specialized knowledge and significant manual effort.

While deep learning offers superior performance in systems and networking, adoption is often hindered by difficulties in understanding and debugging. Explainability aims to bridge this gap by providing insight into the model's decisions. However, existing methods primarily identify the most influential input features, forcing operators to perform extensive manual analysis of low-level signals (e.g., buffer t - 1).

We present CrystalBox, a novel, model-agnostic, posthoc explainability framework for Deep Reinforcement Learning (DRL) controllers in the large family of input-driven environments which includes computer systems. We combine the natural decomposability of reward functions in input-driven environments with the explanatory power of decomposed returns. We propose an efficient algorithm to generate future-based explanations across both discrete and continuous control environments. Using applications such as adaptive bitrate streaming and congestion control, we demonstrate CrystalBox's capability to generate high-fidelity explanations. We further illustrate its higher utility across three practical use cases: contrastive explanations, network observability, and guided reward design, as opposed to prior explainability techniques that identify salient features.

Despite the superior performance of deep learning-based controllers in network applications, their practical adoption is limited due to the difficulty in understanding and trusting them. Existing explainability solutions largely focus on interpreting these controllers by providing insights into the top features used by the model. Although these insights can help reveal an important aspect of the controller, they require operators to deal with low-level features, requiring extensive manual analysis and interpretation.

Rapid advances in machine learning necessitate significant computing power and memory for training, which is accessible only to large corporations today. Small-scale players like academics often only have consumer-grade GPU clusters locally and can afford cloud GPU instances to a limited extent. However, training performance significantly degrades in this multi-cluster setting. In this paper, we identify unique opportunities to accelerate training and propose StellaTrain, a holistic framework that achieves near-optimal training speeds in multi-cloud environments. StellaTrain dynamically adapts a combination of acceleration techniques to minimize time-to-accuracy in model training. StellaTrain introduces novel acceleration techniques such as cache-aware gradient compression and a CPU-based sparse optimizer to maximize GPU utilization and optimize the training pipeline. With the optimized pipeline, StellaTrain holistically determines the training configurations to optimize the total training time. We show that StellaTrain achieves up to 104× speedup over PyTorch DDP in inter-cluster settings by adapting training configurations to fluctuating dynamic network bandwidth. StellaTrain demonstrates that we can cope with the scarce network bandwidth through systematic optimization, achieving up to 257.3× and 78.1× speed-ups on the network bandwidths of 100 Mbps and 500 Mbps, respectively. Finally, StellaTrain enables efficient co-training using on-premises and cloud clusters to reduce costs by 64.5% in conjunction with a reduced training time of 28.9%.

Black swan events are hard-to-predict rare events that can significantly alter the course of our lives. The Internet has played a key role in helping us deal with the coronavirus pandemic, a recent black swan event. However, Internet researchers and operators are mostly blind to another black swan event that poses a direct threat to Internet infrastructure. In this paper, we investigate the impact of solar superstorms that can potentially cause large-scale Internet outages covering the entire globe and lasting several months. We discuss the challenges posed by such activity and currently available mitigation techniques. Using real-world datasets, we analyze the robustness of the current Internet infrastructure and show that submarine cables are at greater risk of failure compared to land cables. Moreover, the US has a higher risk for disconnection compared to Asia. Finally, we lay out steps for improving the Internet's resiliency.

In enterprises, CDNs, and increasingly in edge computing, most data centers have moderate scale. Recent research has developed designs such as expander graphs that are highly efficient compared to large-scale, 3-tier Clos networks, but moderate-scale data centers need to be constructed with standard hardware and protocols familiar to network engineers, and are overwhelmingly built with a leaf-spine architecture. This paper explores whether the performance efficiency that is known to be theoretically possible at large scale can be realized in a practical way for the common leaf-spine data center. First, we find that more efficient topologies indeed exist at moderate scale, showing through simulation and analysis that much of the benefit comes from choosing a 'flat' network that uses one type of switch rather than having separate roles for leafs and spines; indeed, even a simple ring-based topology outperforms leaf-spine for a wide range of traffic scenarios. Second, we design and prototype an efficient routing scheme for flat networks that uses entirely standard hardware and protocols. Our work opens new research directions in topology and routing design that can have significant impact for the most common data centers.

No abstract available.

High throughput is of particular interest in data center and HPC networks. Although myriad network topologies have been proposed, a broad head-to-head comparison across topologies and across traffic patterns is absent, and the right way to compare worst-case throughput performance is a subtle problem. In this paper, we develop a framework to benchmark the throughput of network topologies, using a two-pronged approach. First, we study performance on a variety of synthetic and experimentally-measured traffic matrices (TMs). Second, we show how to measure worst-case throughput by generating a near-worst-case TM for any given topology. We apply the framework to study the performance of these TMs in a wide range of network topologies, revealing insights into the performance of topologies with scaling, robustness of performance across TMs, and the effect of scattered workload placement. Our evaluation code is freely available.

High throughput is a fundamental goal of network design. While myriad network topologies have been proposed to meet this goal, particularly in data center and HPC networking, a consistent and accurate method of evaluating a design's throughput performance and comparing it to past proposals is conspicuously absent. In this work, we develop a framework to benchmark the throughput of network topologies and apply this methodology to reveal insights about network structure. We show that despite being commonly used, cut-based metrics such as bisection bandwidth are the wrong metrics: they yield incorrect conclusions about the throughput performance of networks. We therefore measure flow-based throughput directly and show how to evaluate topologies with nearly-worst-case traffic matrices. We use the flow-based throughput metric to compare the throughput performance of a variety of computer networks. We have made our evaluation framework freely available to facilitate future work on design and evaluation of networks.