Summary Cloud providers place tasks from multiple applications on the same resource pool to improve the resource utilization of the infrastructure. The consequent resource contention has an undesirable effect on latency‐sensitive tasks. In this article, we present Niyama—a resource isolation approach that uses a modified version of deadline scheduling to protect latency‐sensitive tasks from CPU bandwidth contention. Conventionally, deadline scheduling has been used to schedule real‐time tasks with well‐defined deadlines. Therefore, it cannot be used directly when the deadlines are unspecified. In Niyama, we estimate deadlines in intervals and secure bandwidth required for the interval, thereby ensuring optimal job response times. We compare our approach with cgroups: Linux's default resource isolation mechanism used in containers today. Our experiments show that Niyama reduces the average delay in tasks by 3–20 when compared to cgroups. Since Linux's deadline scheduling policy is work‐conserving in nature, there is a small drop in the server‐level CPU utilization when Niyama is used naively. We demonstrate how the use of core reservation and oversubscription in the inter‐node scheduler can be used to offset this drop; our experiments show a 1.3–2.24 decrease in delay in job response time over cgroups while achieving high CPU utilization.

Garbage collection in Java is critical to the performance of Big Data applications. Our study brings forth the causes that invoke the 3 Java garbage collectors for Big Data and non Big Data workloads. We conclude that allocation failure occurs for about 31-99 per cent of the total garbage collector execution time and is the most common cause for garbage collection in Big Data workloads. In case of Dacapo workloads, we observe that the prominent cause for garbage collection is system gc calls which occur for about 40-99 per cent of the total garbage collector execution time.

Summary Distributed applications are popular for heavy workloads where the resources of a single machine are not sufficient. These distributed applications come with many parameters to tune so that cluster resources can be effectively utilized. However, any misconfiguration of the available parameters may result in suboptimal performance of one or more machines in the cluster. These events may go unnoticed or can result in crashes. This problem of misconfigured parameters has no straightforward solution due to the variety of parameters and vastly different workloads being processed. In this article, we propose a methodology for machine learning‐based detection of misconfigurations. We collect data mined from system resource utilization, Hadoop logs, and job‐level metrics to train a model using decision tree and support vector machine. The models are used to identify whether a set of configuration parameters could result in a crash or a slowdown for a specific workload. The approach explained in this article can be extended to other distributed big data applications, such as Spark, Hive, Pig, and so on.

Crosslingual word embeddings developed from multiple parallel corpora help in understanding the relationships between languages and improving the prediction quality of machine translation. However, in low resource languages with complex and agglutinative morphologies, inducing good-quality crosslingual embeddings becomes challenging due to the problem of complex morphological forms and rare words. This is true even for languages that share common linguistic structure. In our work, we have shown that performing a simple morphological segmentation upon the corpora prior to the generation of crosslingual word embeddings for both roots and suffixes greatly improves the prediction quality and captures semantic similarities more effectively. To exhibit this, we have chosen two related languages: Telugu and Kannada of the Dravidian language family. We have also tested our method upon a widely spoken North Indian language, Hindi, belonging to the Indo-European language family, and have observed encouraging results.

Out-of-Vocabulary (OOV) detection and recovery is an important aspect of reducing Word Error Rate (WER) in Automatic Speech Recognition (ASR). In this paper, we evaluate the effect of OOV detection and recovery for a low-resource language on WER. We use a small seed corpus of continuous speech and improve the vocabulary by incorporating the detected OOV words. We use a syllable-model to learn OOV words and augment the word-model with these words leading to improved recognition. Our research investigates the effect on OOV detection and recovery after adding missing syllable sounds in the syllable model using a Text-to-Speech (TTS) system. Our experiments are conducted using 5 hours of continuous speech Kannada corpus. We use an already available Festival TTS for Hindi to generate Kannada speech. Our initial experiments report an improvement in OOV detection due to addition of missing syllable sounds using a cross-lingual TTS system.

Most enterprise data is distributed in multiple relational databases with expert-designed schema. Using traditional single-table machine learning techniques over such data not only incur a computational penalty for converting to a ”flat” form (mega-join), even the human-specified semantic information present in the relations is lost. In this paper, we present a two-phase hierarchical meta-classification algorithm for relational databases with a semantic divide and conquer approach. We propose a recursive, prediction aggregation technique over heterogeneous classifiers applied on individual database tables. A preliminary evaluation on TPCH and UCI benchmarks shows reduced training time without any loss of prediction accuracy.

In typical video delivery systems offering programs on-demand, service should be be nearly immediate and continuous. A video server can provide this type of service by reserving sufficient network and server resources for the duration of playout. Scalability and reduced cost can be achieved using a single channel to serve multiple customers waiting for the same program (referred to as batching). Batching is especially useful during high load periods typically occuring during evening prime time hours. Typical channel allocation algorithms use a greedy, allocate-as-needed policy. Variations in system load can cause these algorithms to suffer poor and unpredictable short-term performance, and non-optimal long term performance. In this paper, we develop a set of realistic workloads, identify the limitations of greedy allocation algorithms, and propose a set of rate-based allocation schemes to solve these limitations. The performance of various video delivery systems are simulated and compared. The rate-based policies are shown to be robust for the workloads examined, and are easy to implement.

In a distributed environment, presentation of structured, composite multimedia information poses new challenges in dealing with variable bandwidth (BW) requirement and synchronization of media data objects. The detailed knowledge of BW requirement obtained by analyzing the document structure can be used to create a prefetch schedule that results in efficient utilization of system resources. A distributed environment consists of various system components that are either dedicated to a client or shared across multiple clients. Shared system components could benefit from Fine Granularity Advanced Reservation (FGAR) of resources based on true BW requirement. Prefetching by utilizing advance knowledge of BW requirement can further improve resource utilization. In this paper, we describe the JINSIL retrieval system that takes into account the available bandwidth and buffer resources and the nature of sharing in each component on the delivery path. It reshapes BW requirement, creates prefetch schedule for efficient resource utilization in each component, and reserves necessary BW and buffer. We also consider good choices for placement of prefetch buffers across various system components.

In a video-on-demand server, resource reservation is required to guarantee continuous delivery. Hence any given storage device (or a striping group treated as a single logical device) can serve only up to a fixed number of client access streams. Each storage device is also limited by the number of video files it can store. For the reasons of availability, incremental growth, and heterogeneity, there may be multiple storage devices in a video server environment. Hence, one or more copies of a particular video may be placed on different storage devices. Since the access rates to different videos are not uniform, there may be load imbalance among the devices. In this paper, we propose a dynamic placement policy (called the Bandwidth to Space Ratio (BSR) Policy) that creates and/or deletes replica of a video, and mixes hot and cold videos so as to make the best use of bandwidth and space of a storage device. The proposed policy is evaluated using a simulation study.

In order to guarantee continuous delivery of a video stream in an on-demand video server environment, a collection of resources (referred to as a logical channel) are reserved in advance. To conserve server resources, multiple client requests for the same video can be batched together and served by a single channel. Increasing the window over which all requests for a particular video are batched results in larger savings in server capacity; however, it also increases the reneging probability of a client. A complication introduced by batching is that if a batched client pauses, a new stream (which may not be immediately available) needs to be started when the client resumes. To provide short response time to resume requests, some channels are set aside and are referred to as contingency channels. To further improve resource utilization, even when a nonbatched client pauses, the channel is released and reacquired upon resume. In this paper, we first develop an analytical model that predicts the reneging probability and expected resume delay, and then use this model to optimally allocate channels for batching, on-demand playback, and contingency. The effectiveness of the proposed policy over a scheme with no contingency channels and no batching is also demonstrated.

In an on-demand video server environment, clients make requests for movies to a centralized video server. Due to the stringent response time requirements, continuous delivery of a video stream to the client has to be guaranteed by reserving sufficient resources required to deliver a stream. Hence there is a hard limit on the number of streams that can be simultaneously delivered by a server. The server can satisfy multiple requests for the same movie using a single disk I/O stream by sending the same data pages to multiple clients (using the multicast facility if present in the system). This can be achieved by batching requests for the same movie that arrive within a short duration of time. In this paper, we consider various policies for selecting the movie to be multicast. The choice of a policy depends very much on the customer waiting time tolerance before reneging. We show that an FCFS policy that schedules the movie with the longest outstanding request can perform better than the MQL policy that chooses the movie with the maximum number of outstanding requests. Additionally, if the user behavior can be influenced by guaranteeing maximum waiting time then it may be beneficial to pre-allocate a fixed number of streams for popular movies. Finally, we demonstrate using empirical distribution for movie requests, that a substantial reduction (of the order of 60%) in required server capacity can be achieved by batching.