Traffic signal control is an important problem in urban mobility with a significant potential for economic and environmental impact. While there is a growing interest in Reinforcement Learning (RL) for traffic signal control, the work so far has focussed on learning through simulations which could lead to inaccuracies due to simplifying assumptions. Instead, real experience data on traffic is available and could be exploited at minimal costs. Recent progress in offline or batch RL has enabled just that. Model-based offline RL methods, in particular, have been shown to generalize from the experience data much better than others.

Organizations are collecting increasingly large amounts of data for data-driven decision making. These data are often dumped into a centralized repository, e.g., a data lake, consisting of thousands of structured and unstructured datasets. Perversely, such mixture makes the problem of discovering tables or documents that are relevant to a user's query very challenging. Despite the recent efforts in data discovery , the problem remains widely open especially in the two fronts of (1) discovering relationships and relatedness across structured and unstructured datasets-where existing techniques suffer from either scalability, being customized for a specific problem type (e.g., entity matching or data integration), or demolishing the structural properties on its way, and (2) developing a holistic system for integrating various similarity measurements and sketches in an effective way to boost the discovery accuracy. In this paper, we propose a new data discovery system, named CMDL, for addressing these two limitations. CMDL supports the data discovery process over both structured and unstructured data while retaining the structural properties of tables. As a result, CMDL is the only system to date that empowers end-users to seamlessly pipeline the discovery tasks across the two modalities. We propose a novel multi-modal embedding representation that captures the similarities between text documents and tabular columns. The model training relies on labeled datasets generated though weak supervision , and thus the system is domain agnostic and easily generalizable. We evaluate CMDL on three real-world data lakes with diverse applications and show that our system is significantly more effective for cross-modality discovery compared to the search-based baseline techniques. Moreover, CMDL is more accurate and robust to different data types and distributions compared to the state-of-the-art systems that are limited to only the structured datasets.

This paper introduces WIRE that manages resources for the DAG-based workflows on IaaS clouds. WIRE predicts and plans resources over the MAPE (Monitor-Analyze-Plan-Execute) loops to: 1) Estimate task performance with online data, 2) Conduct simulations to predict the upcoming loads based on online estimates and workflow DAGs, 3) Apply a resource-steering policy to size cloud instance pools for the maximal parallelism that is consistent with low cost. We implement WIRE on Pegasus WMS/HTCondor and evaluate its performance on the ExoGENI network cloud. The results show that WIRE attains low resource cost with the performance that is typically within a factor of two of optimal.

There is a lot of interest today in building autonomous (or, self-driving) data processing systems. An emerging school of thought is to leverage AI-driven "black box" algorithms for this purpose. In this paper, we present a contrarian view. We study the problem of autotuning the memory allocation for applications running on modern distributed data processing systems. We show that an empirically-driven "white-box" algorithm, called RelM, that we have developed provides a close-to-optimal tuning at a fraction of the overheads compared to state-of-the-art AI-driven "black box" algorithms, namely, Bayesian Optimization (BO) and Deep Distributed Policy Gradient (DDPG). The main reason for RelM's superior performance is that the memory management in modern memory-based data analytics systems is an interplay of algorithms at multiple levels: (i) at the resource-management level across various containers allocated by resource managers like Kubernetes and YARN, (ii) at the container level among the OS, pods, and processes such as the Java Virtual Machine (JVM), (iii) at the application level for caching, aggregation, data shuffles, and application data structures, and (iv) at the JVM level across various pools such as the Young and Old Generation. RelM understands these interactions and uses them in building an analytical solution to autotune the memory management knobs. In another contribution, called Guided-BO (GBO), we use RelM's analytical models to speed up BO. Through an evaluation based on Apache Spark, we showcase that the RelM's recommendations are significantly better than what commonly-used Spark deployments provide, and are close to the ones obtained by brute-force exploration; while GBO provides optimality guarantees for a higher, but still significantly lower cost overhead compared to the state-of-the-art AI-driven policies.

Systems for processing big data---e.g., Hadoop, Spark, and massively parallel databases---need to run workloads on behalf of multiple tenants simultaneously. The abundant disk-based storage in these systems is usually complemented by a smaller, but much faster, cache. Cache is a precious resource: Tenants who get to use the cache can see two orders of magnitude performance improvement. Cache is also a limited and hence shared resource: Unlike a resource like a CPU core which can be used by only one tenant at a time, a cached data item can be accessed by multiple tenants at the same time. Cache, therefore, has to be shared by a multi-tenancy-aware policy across tenants, each having a unique set of priorities and workload characteristics.

Allocation and usage of memory in modern data-processing platforms is based on an interplay of algorithms at multiple levels: (i) at the resource-management level across containers allocated by resource managers like Mesos and Yarn, (ii) at the container level among the OS and processes such as the Java Virtual Machine (JVM), (iii) at the framework level for caching, aggregation, data shuffles, and application data structures, and (iv) at the JVM level across various pools such as the Young and Old Generation as well as the heap versus off-heap. We use Thoth, a data-driven platform for multi-system cluster management, to build a deep understanding of different interplays in memory management options. Through multiple memory management apps built in Thoth, we demonstrate how Thoth can deal with multiple levels of memory management as well as multi-tenant nature of clusters.

Following the 'no one size fits all' philosophy, active research in big data platforms is focusing on creating an environment for multiple 'one-size' systems to co-exist and cooperate in the same cluster. Consequently, it has now become imperative to provide an integrated management solution that provides a database-centric view of the underlying multi-system environment. We outline the proposal of DBMS + , a database management platform over multiple 'one-size' systems. Our prototype implementation of DBMS + , called Thoth , adaptively chooses a best-fit system based on application requirements. In this demonstration, we propose to showcase Thoth DM , a data management framework for Thoth which consists of a data collection pipeline utility, data consolidation and dispatcher module, and a warehouse for storing this data. We further introduce the notion of apps ; an app is a utility that registers with Thoth DM and interfaces with its warehouse to provide core database management functionalities like dynamic provisioning of resources, designing a multi-system-aware optimizer, tuning of configuration parameters on each system, data storage, and layout schemes. We will demonstrate Thoth DM in action over Hive, Hadoop, Shark, Spark, and the Hadoop Distributed File System. This demonstration will focus on the following apps : (i) Dashboard for administration and control that will let the audience monitor and visualize a database-centric view of the multi-system cluster, and (ii) Data Layout Recommender app will allow searching for the optimal data layout in the multi-system setting.

Database engines often consume significant power during query processing activities, motivating researchers to investigate the redesign of their internals to minimize these overheads. While the prior literature has dealt exclusively with average power considerations, our focus here is on peak power consumption. We begin by profiling the peak power behavior of a representative suite of popular commercial database engines in benchmark query processing environments, and demonstrate that their consumption can often be substantial. Then, we develop a pipeline-based model of query execution plans that lends itself to accurately estimating peak power consumption, suggesting its gainful employment in server design and capacity planning. More potently, given a space of competing plan choices, it could help identify plans with attractive tradeoffs between peak-power and time-efficiency considerations, and we present sample instances of such tradeoffs. Finally, we discuss extensions of our modeling approach to inductive pipelines and multi-query workloads.