arXiv subjects

Gil Einziger

Publications and source records attributed to Gil Einziger.

31 records · Page 2Linked to original sources

arXiv2019

Verifying Robustness of Gradient Boosted Models

Gradient boosted models are a fundamental machine learning technique. Robustness to small perturbations of the input is an important quality measure for machine learning models, but the literature lacks a method to prove the robustness of gradient boosted models. This work introduces VeriGB, a tool for quantifying the robustness of gradient boosted models. VeriGB encodes the model and the robustness property as an SMT formula, which enables state of the art verification tools to prove the model's robustness. We extensively evaluate VeriGB on publicly available datasets and demonstrate a capability for verifying large models. Finally, we show that some model configurations tend to be inherently more robust than others.

cs.LG↗

arXiv2018

Memento: Making Sliding Windows Efficient for Heavy Hitters

Cloud operators require real-time identification of Heavy Hitters (HH) and Hierarchical Heavy Hitters (HHH) for applications such as load balancing, traffic engineering, and attack mitigation. However, existing techniques are slow in detecting new heavy hitters. In this paper, we make the case for identifying heavy hitters through \textit{sliding windows}. Sliding windows detect heavy hitters quicker and more accurately than current methods, but to date had no practical algorithms. Accordingly, we introduce, design and analyze the \textit{Memento} family of sliding window algorithms for the HH and HHH problems in the single-device and network-wide settings. Using extensive evaluations, we show that our single-device solutions attain similar accuracy and are by up to $273\times$ faster than existing window-based techniques. Furthermore, we exemplify our network-wide HHH detection capabilities on a realistic testbed. To that end, we implemented Memento as an open-source extension to the popular HAProxy cloud load-balancer. In our evaluations, using an HTTP flood by 50 subnets, our network-wide approach detected the new subnets faster, and reduced the number of undetected flood requests by up to $37\times$ compared to the alternatives.

cs.NI↗

arXiv2018

Efficient Measurement on Programmable Switches Using Probabilistic Recirculation

Programmable network switches promise flexibility and high throughput, enabling applications such as load balancing and traffic engineering. Network measurement is a fundamental building block for such applications, including tasks such as the identification of heavy hitters (largest flows) or the detection of traffic changes. However, high-throughput packet processing architectures place certain limitations on the programming model, such as restricted branching, limited capability for memory access, and a limited number of processing stages. These limitations restrict the types of measurement algorithms that can run on programmable switches. In this paper, we focus on the RMT programmable high-throughput switch architecture, and carefully examine its constraints on designing measurement algorithms. We demonstrate our findings while solving the heavy hitter problem. We introduce PRECISION, an algorithm that uses \emph{Probabilistic Recirculation} to find top flows on a programmable switch. By recirculating a small fraction of packets, PRECISION simplifies the access to stateful memory to conform with RMT limitations and achieves higher accuracy than previous heavy hitter detection algorithms that avoid recirculation. We also analyze the effect of each architectural constraint on the measurement accuracy and provide insights for measurement algorithm designers.

cs.NI↗

arXiv2018

Give Me Some Slack: Efficient Network Measurements

Many networking applications require timely access to recent network measurements, which can be captured using a sliding window model. Maintaining such measurements is a challenging task due to the fast line speed and scarcity of fast memory in routers. In this work, we study the impact of allowing \emph{slack} in the window size on the asymptotic requirements of sliding window problems. That is, the algorithm can dynamically adjust the window size between $W$ and $W(1+τ)$ where $τ$ is a small positive parameter. We demonstrate this model's attractiveness by showing that it enables efficient algorithms to problems such as MAX and GENERAL-SUM that require $Ω(W)$ bits even for constant factor approximations in the exact sliding window model. Additionally, for problems that admit sub-linear approximation algorithms such as BASIC-SUMMING and COUNT-DISTINCT, the slack model enables a further asymptotic improvement.

cs.DS↗

arXiv2018

Scheduling Advertisement Delivery in Vehicular Networks

Vehicular users are emerging as a prime market for targeted advertisement, where advertisements (ads) are sent from network points of access to vehicles, and displayed to passengers only if they are relevant to them. In this study, we take the viewpoint of a broker managing the advertisement system, and getting paid every time a relevant ad is displayed to an interested user. The broker selects the ads to broadcast at each point of access so as to maximize its revenue. In this context, we observe that choosing the ads that best fit the users' interest could actually hurt the broker's revenue. In light of this conflict, we present Volfied, an algorithm allowing for conflict-free, near-optimal ad selection with very low computational complexity. Our performance evaluation, carried out through real-world vehicular traces, shows that Volfied increases the broker revenue by up to 70% with provably low computational complexity, compared to state-of-the-art alternatives.

cs.NI↗

arXiv2017

Pay for a Sliding Bloom Filter and Get Counting, Distinct Elements, and Entropy for Free

For many networking applications, recent data is more significant than older data, motivating the need for sliding window solutions. Various capabilities, such as DDoS detection and load balancing, require insights about multiple metrics including Bloom filters, per-flow counting, count distinct and entropy estimation. In this work, we present a unified construction that solves all the above problems in the sliding window model. Our single solution offers a better space to accuracy tradeoff than the state-of-the-art for each of these individual problems! We show this both analytically and by running multiple real Internet backbone and datacenter packet traces.

cs.DS↗

arXiv2017

Fast Flow Volume Estimation

The increasing popularity of jumbo frames means growing variance in the size of packets transmitted in modern networks. Consequently, network monitoring tools must maintain explicit traffic volume statistics rather than settle for packet counting as before. We present constant time algorithms for volume estimations in streams and sliding windows, which are faster than previous work. Our solutions are formally analyzed and are extensively evaluated over multiple real-world packet traces as well as synthetic ones. For streams, we demonstrate a run-time improvement of up to 2.4X compared to the state of the art. On sliding windows, we exhibit a memory reduction of over 100X on all traces and an asymptotic runtime improvement to a constant. Finally, we apply our approach to hierarchical heavy hitters and achieve an empirical 2.4-7X speedup.

cs.DS↗

arXiv2017

Constant Time Updates in Hierarchical Heavy Hitters

Monitoring tasks, such as anomaly and DDoS detection, require identifying frequent flow aggregates based on common IP prefixes. These are known as \emph{hierarchical heavy hitters} (HHH), where the hierarchy is determined based on the type of prefixes of interest in a given application. The per packet complexity of existing HHH algorithms is proportional to the size of the hierarchy, imposing significant overheads. In this paper, we propose a randomized constant time algorithm for HHH. We prove probabilistic precision bounds backed by an empirical evaluation. Using four real Internet packet traces, we demonstrate that our algorithm indeed obtains comparable accuracy and recall as previous works, while running up to 62 times faster. Finally, we extended Open vSwitch (OVS) with our algorithm and showed it is able to handle 13.8 million packets per second. In contrast, incorporating previous works in OVS only obtained 2.5 times lower throughput.

cs.DS↗

arXiv2017

Optimal Elephant Flow Detection

Monitoring the traffic volumes of elephant flows, including the total byte count per flow, is a fundamental capability for online network measurements. We present an asymptotically optimal algorithm for solving this problem in terms of both space and time complexity. This improves on previous approaches, which can only count the number of packets in constant time. We evaluate our work on real packet traces, demonstrating an up to X2.5 speedup compared to the best alternative.

cs.DS↗

arXiv2016

Randomized Admission Policy for Efficient Top-k and Frequency Estimation

Network management protocols often require timely and meaningful insight about per flow network traffic. This paper introduces Randomized Admission Policy (RAP) - a novel algorithm for the frequency and top-k estimation problems, which are fundamental in network monitoring. We demonstrate space reductions compared to the alternatives by a factor of up to 32 on real packet traces and up to 128 on heavy-tailed workloads. For top-k identification, RAP exhibits memory savings by a factor of between 4 and 64 depending on the skew of the workload. These empirical results are backed by formal analysis, indicating the asymptotic space improvement of our probabilistic admission approach. Additionally, we present d-Way RAP, a hardware friendly variant of RAP that empirically maintains its space and accuracy benefits.

cs.DS↗

arXiv2016

ICE Buckets: Improved Counter Estimation for Network Measurement

Measurement capabilities are essential for a variety of network applications, such as load balancing, routing, fairness and intrusion detection. These capabilities require large counter arrays in order to monitor the traffic of all network flows. While commodity SRAM memories are capable of operating at line speed, they are too small to accommodate large counter arrays. Previous works suggested estimators, which trade precision for reduced space. However, in order to accurately estimate the largest counter, these methods compromise the accuracy of the smaller counters. In this work, we present a closed form representation of the optimal estimation function. We then introduce Independent Counter Estimation Buckets (ICE-Buckets), a novel algorithm that improves estimation accuracy for all counters. This is achieved by separating the flows to buckets and configuring the optimal estimation function according to each bucket's counter scale. We prove a tighter upper bound on the relative error and demonstrate an accuracy improvement of up to 57 times on real Internet packet traces.

cs.NI↗

arXiv2016

Efficient Summing over Sliding Windows

This paper considers the problem of maintaining statistic aggregates over the last W elements of a data stream. First, the problem of counting the number of 1's in the last W bits of a binary stream is considered. A lower bound of Ω(1/ε + log W) memory bits for Wε-additive approximations is derived. This is followed by an algorithm whose memory consumption is O(1/ε + log W) bits, indicating that the algorithm is optimal and that the bound is tight. Next, the more general problem of maintaining a sum of the last W integers, each in the range of {0,1,...,R}, is addressed. The paper shows that approximating the sum within an additive error of RWε can also be done using Θ(1/ε + log W) bits for ε=Ω(1/W). For ε=o(1/W), we present a succinct algorithm which uses B(1 + o(1)) bits, where B=Θ(Wlog(1/Wε)) is the derived lower bound. We show that all lower bounds generalize to randomized algorithms as well. All algorithms process new elements and answer queries in O(1) worst-case time.

cs.DS↗

arXiv2015

TinyLFU: A Highly Efficient Cache Admission Policy

This paper proposes to use a frequency based cache admission policy in order to boost the effectiveness of caches subject to skewed access distributions. Given a newly accessed item and an eviction candidate from the cache, our scheme decides, based on the recent access history, whether it is worth admitting the new item into the cache at the expense of the eviction candidate. Realizing this concept is enabled through a novel approximate LFU structure called TinyLFU, which maintains an approximate representation of the access frequency of a large sample of recently accessed items. TinyLFU is very compact and light-weight as it builds upon Bloom filter theory. We study the properties of TinyLFU through simulations of both synthetic workloads as well as multiple real traces from several sources. These simulations demonstrate the performance boost obtained by enhancing various replacement policies with the TinyLFU eviction policy. Also, a new combined replacement and eviction policy scheme nicknamed W-TinyLFU is presented. W-TinyLFU is demonstrated to obtain equal or better hit-ratios than other state of the art replacement policies on these traces. It is the only scheme to obtain such good results on all traces.

cs.OS↗