Parimal Parag

Indian Institute of Science Bangalore | IISC · Department of Electrical and Communication Engineering

This paper investigates the $(k,k)$ fork-join scheduling scheme on a system of $n$ parallel servers comprising both slow and fast servers. Tasks arriving in the system are divided into $k$ sub-tasks and assigned to a random set of $k$ servers, where each task can be assigned independently to a distinct slow or fast server with selection pro...

We study optimal pricing in a single server queue when the customers valuation of service depends on their waiting time. In particular, we consider a very general model, where the customer valuations are random and are sampled from a distribution that depends on the queue length. The goal of the service provider is to set dynamic state dependent pr...

We consider (k, k) fork-join scheduling on a large number (say, N) of parallel servers with two sets of heterogeneous rates. An incoming task is split into k sub-tasks and dispatched to k servers according to a probabilistic selection policy, with parameter ps being the selection probability of slower servers. Mean task completion time admits an in...

We consider the setting where a master wants to run a distributed stochastic gradient descent (SGD) algorithm on $n$ workers, each having a subset of the data. Distributed SGD may suffer from the effect of stragglers, i.e., slow or unresponsive workers who cause delays. One solution studied in the literature is to wait at each iteration for the res...

We consider the setting of a distributed storage system where a single file is subdivided into smaller fragments of same size which are then replicated with a common replication factor across servers of identical cache size. An incoming file download request is sent to all the servers, and the download is completed whenever the request gathers all...

We study optimal service pricing in server farms where customers arrive according to a renewal process and have independent and identical (i.i.d. ) exponential service times and i.i.d. valuations of the service. The service provider charges a time varying service fee aiming at maximizing its revenue rate. The customers that find free servers and se...

Data is replicated and stored redundantly over multiple servers for availability in distributed databases. We focus on databases with frequent reads and writes, where both read and write latencies are important. This is in contrast to databases designed primarily for either read or write applications. Redundancy has contrasting effects on read and...

We consider energy minimization for data-intensive applications run on large number of servers, for given performance guarantees. We consider a system, where each incoming application is sent to a set of servers, and is considered to be completed if a subset of them finish serving it. We consider a simple case when each server core has two speed le...

The ability of a P2P network to scale its throughput up in proportion to the arrival rate of peers has recently been shown to be crucially dependent on the chunk sharing policy employed. Some policies can result in low frequencies of a particular chunk, known as the missing chunk syndrome, which can dramatically reduce throughput and lead to instab...

Given the unpredictable nature of the nodes in distributed computing systems, some of the tasks can be significantly delayed. Such delayed tasks are called stragglers. Straggler mitigation can be achieved by redundant computation. In maximum distance separable (MDS) redundancy method, a task is divided into $k$ subtasks which are encoded to $n$...

Samples from a high-dimensional first-order auto-regressive process generated by an independently and identically distributed random innovation sequence are observed by a sender which can communicate only finitely many bits per unit time to a receiver. The receiver seeks to form an estimate of the process value at every time instant in real-time. W...

Timely reception of information is of utmost importance in many cyber-physical systems. We study timely status update schemes for a single source. We consider a scheme which exploits the temporal correlation in source messages to send differential information to the receiver, over an unreliable channel. In this scheme, the source sends the true sta...

Dispatching policies such as the join shortest queue (JSQ), join smallest work (JSW) and their power of two variants are used in load balancing systems where the instantaneous queue length or workload information at all queues or a subset of them can be queried. In situations where the dispatcher has an associated memory, one can minimize this quer...

We consider the setting of distributed storage system where a single file is subdivided into smaller fragments of same size which are then replicated with a common replication factor across servers of identical cache size. An incoming file download request is sent to all the servers, and the download is completed whenever request gathers all the fr...

Real-time decision making relies on the availability of accurate data and, therefore, delivering status updates in a timely fashion is of paramount importance. The topic of real-time status updates has received much attention in recent years. This article contributes new results to this research area by studying the interplay between average timeli...

A transmitter observing a sequence of independent and identically distributed random variables seeks to keep a receiver updated about its latest observations. The receiver need not be apprised about each symbol seen by the transmitter, but needs to output a symbol at each time instant $t$ . If at time $t$ the receiver outputs the symbol seen by...

Samples from a high-dimensional AR[1] process are observed by a sender which can communicate only finitely many bits per unit time to a receiver. The receiver seeks to form an estimate of the process value at every time instant in real-time. We consider a time-slotted communication model in a slow-sampling regime where multiple communication slots...

The performance of large-scale distributed compute systems is adversely impacted by stragglers when the execution time of a job is uncertain. To manage stragglers, we consider a multi-fork approach for job scheduling, where additional parallel servers are added at forking instants. In terms of the forking instants and the number of additional serve...

We consider the setting where a master wants to run a distributed stochastic gradient descent (SGD) algorithm on n workers each having a subset of the data. Distributed SGD may suffer from the effect of stragglers, i.e., slow or unresponsive workers who cause delays. One solution studied in the literature is to wait at each iteration for the respon...

We consider the setting where a master wants to run a distributed stochastic gradient descent (SGD) algorithm on $n$ workers each having a subset of the data. Distributed SGD may suffer from the effect of stragglers, i.e., slow or unresponsive workers who cause delays. One solution studied in the literature is to wait at each iteration for the resp...

The performance of large-scale distributed compute systems is adversely impacted by stragglers when the execution time of a job is uncertain. To manage stragglers, we consider a multi-fork approach for job scheduling, where additional parallel servers are added at forking instants. In terms of the forking instants and the number of additional serve...

Modern communication and computation systems often consist of large networks of unreliable nodes. Still, it is well known that such systems can provide aggregate reliability via redundancy. While duplication may increase the load on a system, it can lead to significant performance improvement when combined with the judicious management of extra sys...

For timely sensor update, the traditional approach is to send new information at every available opportunity. Recent research has shown that with limited receiver feedback, sensors can improve the update timeliness by transmitting differential information for slowly varying correlated sources. One can elect to transmit either the actual or the diff...

The ability of a P2P network to scale its throughput up in proportion to the arrival rate of peers has recently been shown to be crucially dependent on the chunk sharing policy employed. Some policies can result in low frequencies of a particular chunk, known as the missing chunk syndrome, which can dramatically reduce throughput and lead to instab...

This article focuses on the design and evaluation of pragmatic schemes for delay-sensitive communication. Specifically, this contribution studies the operation of data links that employ incremental redundancy as a means to shield information bits from the degradation associated with unreliable channels. While this inquiry puts forth a general metho...

We consider status updates of a physical process over an unreliable channel. In this setting, one may not be able to reliably transmit the current state at all times. Instead, one is interested in the timeliness of the accurately received information. This is a setting for several cyber-physical system applications that require real-time monitoring...

A transmitter observing a sequence of independent and identically distributed random variables seeks to keep a receiver updated about its latest observations. The receiver need not be apprised about each symbol seen by the transmitter, but needs to output a symbol at each time instant $t$. If at time $t$ the receiver outputs the symbol seen by the...

As sensing and instrumentation play an increasingly important role in systems controlled over wired and wireless networks, the need to better understand delay-sensitive communication becomes a prime issue. Along these lines, this article studies the operation of data links that employ incremental redundancy as a practical means to protect informati...

We consider the setting of a Master server, M, who possesses confidential data (e.g., personal, genomic or medical data) and wants to run intensive computations on it, as part of a machine learning algorithm for example. The Master wants to distribute these computations to untrusted workers who have volunteered or are incentivized to help with this...

Modern communication and computation systems consist of large networks of unreliable nodes. Yet, it is well known that such systems can provide aggregate reliability via information redundancy, duplicating paths, or replicating computations. While redundancy may increase the load on a system, it can also lead to major performance improvements throu...

For many applications in sensor networks and cyber-physical systems, receiving timely information is of utmost importance. In this article, we study data transmission schemes for a single source, sending periodic updates to a receiver through an unreliable channel. We consider two schemes that exploit the temporal correlation in the source messages...

We consider the setting of a master server who possesses confidential data (genomic, medical data, etc.) and wants to run intensive computations on it, as part of a machine learning algorithm for example. The master wants to distribute these computations to untrusted workers who have volunteered or are incentivized to help with this task. However,...

This paper considers the relationship between code-rate selection and queueing performance for communication systems subject to time-varying channel conditions. While error-correcting codes offer protection against channel uncertainties, there exists a natural tradeoff between the enhanced protection of low-rate codes and the rate penalty imposed b...

Quantization is intrinsic to several data acquisition systems. This process is especially important in distributed settings, where observations must first be compressed before they are disseminated. There have been many practical successes in the area of quantization, including the acclaimed Lloyd-Max algorithm. This article adopts a different pers...

The traditional formulation of the total value of information transfer is a multi-commodity flow problem. Each data source is seen as generating a commodity along a fixed route, and the objective is to maximize the total system throughput under some concept of fairness, subject to capacity constraints of the links used. This problem is well studied...

The popularity of Peer-to-Peer (P2P) file sharing has resulted in large flows between different ISPs, which imposes significant transit fees on the ISPs in whose domains the communicating peers are located. The fundamental tradeoff faced by a peer-swarm is between free, yet delayed content exchange between intra-domain peers, and inter-domain commu...

The rapid increase of content delivery over the Internet has led to the proliferation of content distribution networks (CDNs). Management of CDNs requires algorithms for request routing, content placement, and eviction in such a way that user delays are small. We abstract the system of frontend source nodes and backend caches of the CDN in the like...

This paper considers the queueing performance of a system that transmits coded data over a time-varying erasure channel. In our model, the queue length and channel state together form a Markov chain that depends on the system parameters. This gives a framework that allows a rigorous analysis of the queue as a function of the code rate. Most prior w...

This paper considers the queueing performance of a system that transmits coded data over a time-varying erasure channel. In our model, the queue length and channel state together form a Markov chain that depends on the system parameters. This gives a framework that allows a rigorous analysis of the queue as a function of the code rate. Most prior w...

Network coding has gained significant attention in recent years as a means to improve throughput, especially in multicast scenarios. These capacity gains are achieved by combining packets algebraically at various points in the network, thereby alleviating local congestion at the nodes. The benefits of network coding are greatest when the network is...

The traditional formulation of the total value of information transfer is a multi-commodity flow problem. Here, each data source is seen as generating a commodity along a fixed route, and the objective is to maximize the total system throughput under some concept of fairness, subject to capacity constraints of the links used. This problem is well s...

This paper considers the relationship between code-rate selection and queueing performance for communication systems with time-varying parameters. While error-correcting codes offer protection against channel unreliability, there is a tradeoff between the enhanced protection of low-rate codes and the increased information transfer of high-rate code...

Network coding has gained significant attention in recent years as a means to improve throughput, especially in multicast scenarios. These capacity gains are achieved by combining packets algebraically at various points in the network, thereby alleviating local congestion at the nodes. The benefits of network coding are greatest when the network is...

A wireless communication system in which multiple users cooperate to transmit information to a common destination is considered. The traffic generated by the users is subject to a stringent quality of service requirement, which is defined in terms of the asymptotic decay-rate of buffer occupancy. The performance of this communication system is anal...

Wireless systems offer a unique mixture of connectivity, flexibility, and freedom. It is therefore not surprising that wireless technology is being embraced with increasing vigor. For real-time applications, user satisfaction is closely linked to quantities such as queue length, packet loss probability, and delay. System performance is therefore re...

We propose a new subcarrier allocation algorithm for Orthogonal Frequency Division Multiple Access (OFDMA) that gives fair allocation of capacity to multiple users with different channel and traffic characteristics. This is achieved by utilizing buffer state information and measured traffic statistics in addition to channel state feedback. Multiuse...