arXiv subjects

Srikanth Kandula

Publications and source records attributed to Srikanth Kandula.

17 recordsLinked to original sources

arXiv2025

Speculative Ad-hoc Querying

Analyzing large datasets requires responsive query execution, but executing SQL queries on massive datasets can be slow. This paper explores whether query execution can begin even before the user has finished typing, allowing results to appear almost instantly. We propose SpeQL, a system that leverages Large Language Models (LLMs) to predict likely queries based on the database schema, the user's past queries, and their incomplete query. Since exact query prediction is infeasible, SpeQL speculates on partial queries in two ways: 1) it predicts the query structure to compile and plan queries in advance, and 2) it precomputes smaller temporary tables that are much smaller than the original database, but are still predicted to contain all information necessary to answer the user's final query. Additionally, SpeQL continuously displays results for speculated queries and subqueries in real time, aiding exploratory analysis. A utility/user study showed that SpeQL improved task completion time, and participants reported that its speculative display of results helped them discover patterns in the data more quickly. In the study, SpeQL improves user's query latency by up to $289\times$ and kept the overhead reasonable, at $\$4$ per hour.

cs.DB↗

arXiv2023

Finding Adversarial Inputs for Heuristics using Multi-level Optimization

Production systems use heuristics because they are faster or scale better than their optimal counterparts. Yet, practitioners are often unaware of the performance gap between a heuristic and the optimum or between two heuristics in realistic scenarios. We present MetaOpt, a system that helps analyze heuristics. Users specify the heuristic and the optimal (or another heuristic) as input, and MetaOpt automatically encodes these efficiently for a solver to find performance gaps and their corresponding adversarial inputs. Its suite of built-in optimizations helps it scale its analysis to practical problem sizes. To show it is versatile, we used MetaOpt to analyze heuristics from three domains (traffic engineering, vector bin packing, and packet scheduling). We found a production traffic engineering heuristic can require 30% more capacity than the optimal to satisfy realistic demands. Based on the patterns in the adversarial inputs MetaOpt produced, we modified the heuristic to reduce its performance gap by 12.5$\times$. We examined adversarial inputs to a vector bin packing heuristic and proved a new lower bound on its performance.

cs.NI↗

arXiv2023

Solving Max-Min Fair Resource Allocations Quickly on Large Graphs

We consider the max-min fair resource allocation problem. The best-known solutions use either a sequence of optimizations or waterfilling, which only applies to a narrow set of cases. These solutions have become a practical bottleneck in WAN traffic engineering and cluster scheduling, especially at larger problem sizes. We improve both approaches: (1) we show how to convert the optimization sequence into a single fast optimization, and (2) we generalize waterfilling to the multi-path case. We empirically show our new algorithms Pareto-dominate prior techniques: they produce faster, fairer, and more efficient allocations. Some of our allocators also have theoretical guarantees: they trade off a bounded amount of unfairness for faster allocation. We have deployed our allocators in Azure's WAN traffic engineering pipeline, where we preserve solution quality and achieve a roughly $3\times$ speedup.

cs.NI↗

arXiv2023

Enhancing Network Management Using Code Generated by Large Language Models

Analyzing network topologies and communication graphs plays a crucial role in contemporary network management. However, the absence of a cohesive approach leads to a challenging learning curve, heightened errors, and inefficiencies. In this paper, we introduce a novel approach to facilitate a natural-language-based network management experience, utilizing large language models (LLMs) to generate task-specific code from natural language queries. This method tackles the challenges of explainability, scalability, and privacy by allowing network operators to inspect the generated code, eliminating the need to share network data with LLMs, and concentrating on application-specific requests combined with general program synthesis techniques. We design and evaluate a prototype system using benchmark applications, showcasing high accuracy, cost-effectiveness, and the potential for further enhancements using complementary program synthesis techniques.

cs.NI↗

arXiv2023

Enhancing Network Failure Mitigation with Performance-Aware Ranking

Cloud providers install mitigations to reduce the impact of network failures within their datacenters. Existing network mitigation systems rely on simple local criteria or global proxy metrics to determine the best action. In this paper, we show that we can support a broader range of actions and select more effective mitigations by directly optimizing end-to-end flow-level metrics and analyzing actions holistically. To achieve this, we develop novel techniques to quickly estimate the impact of different mitigations and rank them with high fidelity. Our results on incidents from a large cloud provider show orders of magnitude improvements in flow completion time and throughput. We also show our approach scales to large datacenters.

cs.NI↗

arXiv2023

Rethinking Machine Learning Collective Communication as a Multi-Commodity Flow Problem

We show communication schedulers' recent work proposed for ML collectives does not scale to the increasing problem sizes that arise from training larger models. These works also often produce suboptimal schedules. We make a connection with similar problems in traffic engineering and propose a new method, TECCL, that finds better quality schedules (e.g., finishes collectives faster and/or while sending fewer bytes) and does so more quickly on larger topologies. We present results on many different GPU topologies that show substantial improvement over the state-of-the-art.

cs.NI↗

arXiv2023

A Deep Learning Perspective on Network Routing

Routing is, arguably, the most fundamental task in computer networking, and the most extensively studied one. A key challenge for routing in real-world environments is the need to contend with uncertainty about future traffic demands. We present a new approach to routing under demand uncertainty: tackling this challenge as stochastic optimization, and employing deep learning to learn complex patterns in traffic demands. We show that our method provably converges to the global optimum in well-studied theoretical models of multicommodity flow. We exemplify the practical usefulness of our approach by zooming in on the real-world challenge of traffic engineering (TE) on wide-area networks (WANs). Our extensive empirical evaluation on real-world traffic and network topologies establishes that our approach's TE quality almost matches that of an (infeasible) omniscient oracle, outperforming previously proposed approaches, and also substantially lowers runtimes.

cs.NI↗

arXiv2021

Energy-Proportional Data Center Network Architecture Through OS, Switch and Laser Co-design

Optical interconnects are already the dominant technology in large-scale data center networks. However, the high optical loss of many optical components coupled with the low efficiency of laser sources result in high aggregate power requirements for the thousands of optical transceivers used by these networks. As optical interconnects stay always on even as traffic demands ebb and flow, most of this power is wasted. We present LC/DC, a data center network system architecture in which the operating system, the switch, and the optical components are co-designed to achieve energy proportionality. LC/DC capitalizes on the path divergence of data center networks to turn on and off redundant paths according to traffic demand, while maintaining full connectivity. Turning off redundant paths allows the optical transceivers and their electronic drivers to power down and save energy. Maintaining full connectivity hides the laser turn-on delay. At the node layer, intercepting send requests within the OS allows for the NIC's laser turn-on delay to be fully overlapped with TCP/IP packet processing, and thus egress links can remain powered off until needed with zero performance penalty. We demonstrate the feasibility of LC/DC by i) implementing the necessary modifications in the Linux kernel and device drivers, ii) implementing a 10Gbit/s FPGA switch, and iii) performing physical experiments with optical devices and circuit simulations. Our results on university data center traces and models of Facebook and Microsoft data center traffic show that LC/DC saves on average 60% of the optical transceivers power (68% max) at the cost of 6% higher packet delay.

cs.NI↗

arXiv2021

Solving Large-Scale Granular Resource Allocation Problems Efficiently with POP

Resource allocation problems in many computer systems can be formulated as mathematical optimization problems. However, finding exact solutions to these problems using off-the-shelf solvers is often intractable for large problem sizes with tight SLAs, leading system designers to rely on cheap, heuristic algorithms. We observe, however, that many allocation problems are granular: they consist of a large number of clients and resources, each client requests a small fraction of the total number of resources, and clients can interchangeably use different resources. For these problems, we propose an alternative approach that reuses the original optimization problem formulation and leads to better allocations than domain-specific heuristics. Our technique, Partitioned Optimization Problems (POP), randomly splits the problem into smaller problems (with a subset of the clients and resources in the system) and coalesces the resulting sub-allocations into a global allocation for all clients. We provide theoretical and empirical evidence as to why random partitioning works well. In our experiments, POP achieves allocations within 1.5% of the optimal with orders-of-magnitude improvements in runtime compared to existing systems for cluster scheduling, traffic engineering, and load balancing.

cs.DC↗

arXiv2021

Towards a Cost vs. Quality Sweet Spot for Monitoring Networks

Continuously monitoring a wide variety of performance and fault metrics has become a crucial part of operating large-scale datacenter networks. In this work, we ask whether we can reduce the costs to monitor -- in terms of collection, storage and analysis -- by judiciously controlling how much and which measurements we collect. By positing that we can treat almost all measured signals as sampled time-series, we show that we can use signal processing techniques such as the Nyquist-Shannon theorem to avoid wasteful data collection. We show that large savings appear possible by analyzing tens of popular measurements from a production datacenter network. We also discuss the technical challenges that must be solved when applying these techniques in practice.

cs.NI↗

arXiv2021

Q-error Bounds of Random Uniform Sampling for Cardinality Estimation

Random uniform sampling has been studied in various statistical tasks but few of them have covered the Q-error metric for cardinality estimation (CE). In this paper, we analyze the confidence intervals of random uniform sampling with and without replacement for single-table CE. Results indicate that the upper Q-error bound depends on the sample size and true cardinality. Our bound gives a rule-of-thumb for how large a sample should be kept for single-table CE.

math.ST↗

arXiv2020

Approximate Partition Selection for Big-Data Workloads using Summary Statistics

Many big-data clusters store data in large partitions that support access at a coarse, partition-level granularity. As a result, approximate query processing via row-level sampling is inefficient, often requiring reads of many partitions. In this work, we seek to answer queries quickly and approximately by reading a subset of the data partitions and combining partial answers in a weighted manner without modifying the data layout. We illustrate how to efficiently perform this query processing using a set of pre-computed summary statistics, which inform the choice of partitions and weights. We develop novel means of using the statistics to assess the similarity and importance of partitions. Our experiments on several datasets and data layouts demonstrate that to achieve the same relative error compared to uniform partition sampling, our techniques offer from 2.7$\times$ to $70\times$ reduction in the number of partitions read, and the statistics stored per partition require fewer than 100KB.

cs.DB↗

arXiv2019

Efficient Inter-Datacenter Bulk Transfers with Mixed Completion Time Objectives

Bulk transfers from one to multiple datacenters can have many different completion time objectives ranging from quickly replicating some $k$ copies to minimizing the time by which the last destination receives a full replica. We design an SDN-style wide-area traffic scheduler that optimizes different completion time objectives for various requests. The scheduler builds, for each bulk transfer, one or more multicast forwarding trees which preferentially use lightly loaded network links. Multiple multicast trees are used per bulk transfer to insulate destinations that have higher available bandwidth and can hence finish quickly from congested destinations. These decisions--how many trees to construct and which receivers to serve using a given tree--result from an optimization problem that minimizes a weighted sum of transfers' completion time objectives and their bandwidth consumption. Results from simulations and emulations on Mininet show that our scheduler, Iris, can improve different completion time objectives by about $2.5\times$.

cs.DC↗

arXiv2018

Fast and Efficient Bulk Multicasting over Dedicated Inter-Datacenter Networks

Several organizations have built multiple datacenters connected via dedicated wide area networks over which large inter-datacenter transfers take place. This includes tremendous volumes of bulk multicast traffic generated as a result of data and content replication. Although one can perform these transfers using a single multicast forwarding tree, that can lead to poor performance as the slowest receiver on each tree dictates the completion time for all receivers. Using multiple trees per transfer each connected to a subset of receivers alleviates this concern. The choice of multicast trees also determines the total bandwidth usage. To further improve the performance, bandwidth over dedicated inter-datacenter networks can be carved for different multicast trees over specific time periods to avoid congestion and minimize the average receiver completion times. In this paper, we break this problem into the three sub-problems of partitioning, tree selection, and rate allocation. We present an algorithm called QuickCast which is computationally fast and allows us to significantly speed up multiple receivers per bulk multicast transfer with control over extra bandwidth consumption. We evaluate QuickCast against a variety of synthetic and real traffic patterns as well as real WAN topologies. Compared to performing bulk multicast transfers as separate unicast transfers, QuickCast achieves up to $3.64\times$ reduction in mean completion times while at the same time using $0.71\times$ the bandwidth. Also, QuickCast allows the top $50\%$ of receivers to complete between $3\times$ to $35\times$ faster on average compared with when a single forwarding multicast tree is used for data delivery.

cs.NI↗

arXiv2018

QuickCast: Fast and Efficient Inter-Datacenter Transfers using Forwarding Tree Cohorts

Large inter-datacenter transfers are crucial for cloud service efficiency and are increasingly used by organizations that have dedicated wide area networks between datacenters. A recent work uses multicast forwarding trees to reduce the bandwidth needs and improve completion times of point-to-multipoint transfers. Using a single forwarding tree per transfer, however, leads to poor performance because the slowest receiver dictates the completion time for all receivers. Using multiple forwarding trees per transfer alleviates this concern--the average receiver could finish early; however, if done naively, bandwidth usage would also increase and it is apriori unclear how best to partition receivers, how to construct the multiple trees and how to determine the rate and schedule of flows on these trees. This paper presents QuickCast, a first solution to these problems. Using simulations on real-world network topologies, we see that QuickCast can speed up the average receiver's completion time by as much as $10\times$ while only using $1.04\times$ more bandwidth; further, the completion time for all receivers also improves by as much as $1.6\times$ faster at high loads.

cs.NI↗

arXiv2017

DCCast: Efficient Point to Multipoint Transfers Across Datacenters

Using multiple datacenters allows for higher availability, load balancing and reduced latency to customers of cloud services. To distribute multiple copies of data, cloud providers depend on inter-datacenter WANs that ought to be used efficiently considering their limited capacity and the ever-increasing data demands. In this paper, we focus on applications that transfer objects from one datacenter to several datacenters over dedicated inter-datacenter networks. We present DCCast, a centralized Point to Multi-Point (P2MP) algorithm that uses forwarding trees to efficiently deliver an object from a source datacenter to required destination datacenters. With low computational overhead, DCCast selects forwarding trees that minimize bandwidth usage and balance load across all links. With simulation experiments on Google's GScale network, we show that DCCast can reduce total bandwidth usage and tail Transfer Completion Times (TCT) by up to $50\%$ compared to delivering the same objects via independent point-to-point (P2P) transfers.

cs.NI↗

arXiv2016

Do the Hard Stuff First: Scheduling Dependent Computations in Data-Analytics Clusters

We present a scheduler that improves cluster utilization and job completion times by packing tasks having multi-resource requirements and inter-dependencies. While the problem is algorithmically very hard, we achieve near-optimality on the job DAGs that appear in production clusters at a large enterprise and in benchmarks such as TPC-DS. A key insight is that carefully handling the long-running tasks and those with tough-to-pack resource needs will produce good-enough schedules. However, which subset of tasks to treat carefully is not clear (and intractable to discover). Hence, we offer a search procedure that evaluates various possibilities and outputs a preferred schedule order over tasks. An online component enforces the schedule orders desired by the various jobs running on the cluster. In addition, it packs tasks, overbooks the fungible resources and guarantees bounded unfairness for a variety of desirable fairness schemes. Relative to the state-of-the art schedulers, we speed up 50% of the jobs by over 30% each.

cs.DC↗