Wojciech M. Golab

Policy-driven automated reconfiguration for performance management in WDM optical networks

A key feature of optical networks based on WDM technology is the ability to optimize the configuration of optimal resources (i.e., wavelengths) with respect to a particular traffic demand. In the broadcast architecture, this involves the assignment of wavelengths to logical links, while in the optically switched architecture it additionally involves the routing of all-optical data paths known as lightpaths. This survey article is concerned with the problem of automatically updating the configuration of an optical network to accommodate changes in traffic demand, which entails making a reconfiguration policy decision, selecting a new configuration and migrating from the current to the new configuration. Existing solutions are classified according to their algorithmic properties, and compared on the basis performance, computational cost, and flexibility. Finally, open problems and research directions are discussed.

Lightpaths on demand: a Web-services-based management system

User-controlled optical networks play a key role in supporting electronic transfer of the enormous volumes of data generated in emerging e-science experiments. The ability of users to manage their own resources enables provisioning of bandwidth-guaranteed tunnels on demand without the costs associated with conventional managed services offered by network providers. However, building high-performance user-controlled networks has only become feasible in the last few years, as trends in the telecommunications industry have made it possible for users to purchase installed optical fiber and light it using their own premises equipment. Consequently, suitable network management technologies have not yet evolved. In particular, there is presently no means for users to easily provision bandwidth-guaranteed tunnels across multiple independent management domains. In this article we present a user-controlled lightpath management system that addresses this problem. We begin by reviewing the high-level functionality of the system. Then we examine the software architecture. Finally, we discuss design challenges faced while building the system and propose future extensions.

Grid-Controlled Lightpaths for High Performance Grid Applications

Grid applications call for high performance networking support. One attractive solution is to deploy Grids over optical networks. However, resource management in optical domains is traditionally very rigid and cannot successfully meet the requirements of Grid applications, such as flexible provisioning and configuration. In this paper, we present a customizable resource management solution for optical networks where users can create lightpaths on demand and manage their own network resources. Thanks to a Grid-centric system architecture, lightpath resources can be shared among users and easily integrated with data and computation Grids.

Client-Centric Benchmarking of Eventual Consistency for Cloud Storage Systems

Eventually-consistent key-value storage systems sacrifice the ACID semantics of conventional databases to achieve superior latency and availability. However, this means that client applications, and hence end-users, can be exposed to stale data. The degree of staleness observed depends on various tuning knobs set by application developers (customers of key-value stores) and system administrators (providers of key-value stores). Both parties must be cognizant of how these tuning knobs affect the consistency observed by client applications in the interest of both providing the best end-user experience and maximizing revenues for storage providers. Quantifying consistency in a meaningful way is a critical step toward both understanding what clients actually observe, and supporting consistency-aware service level agreements (SLAs) in next generation storage systems. This paper proposes a novel consistency metric called Gamma that captures client-observed consistency. This metric provides quantitative answers to questions regarding observed consistency anomalies, such as how often they occur and how bad they are when they do occur. We argue that Gamma is more useful and accurate than existing metrics. We also apply Gamma to benchmark the popular Cassandra key-value store. Our experiments demonstrate that Gamma is sensitive to both the workload and client-level tuning knobs, and is preferable to existing techniques which focus on worst-case behavior.

Client-centric benchmarking of eventual consistency for cloud storage systems

Eventually-consistent key-value storage systems sacrifice the ACID semantics of conventional databases to achieve superior latency and availability. However, this means that client applications, and hence end-users, can be exposed to stale data. The degree of staleness observed depends on various tuning knobs set by application developers (customers of key-value stores) and system administrators (providers of key-value stores). Both parties must be cognizant of how these tuning knobs affect the consistency observed by client applications in the interest of both providing the best end-user experience and maximizing revenues for storage providers. Quantifying consistency in a meaningful way is a critical step toward both understanding what clients actually observe, and supporting consistency-aware service level agreements (SLAs) in next generation storage systems. This paper proposes a novel consistency metric called Gamma that captures client-observed consistency. This metric provides quantitative answers to questions regarding observed consistency anomalies, such as how often they occur and how bad they are when they do occur. We argue that Gamma is more useful and accurate than existing metrics. We also apply Gamma to benchmark the popular Cassandra key-value store. Our experiments demonstrate that Gamma is sensitive to both the workload and client-level tuning knobs, and is preferable to existing techniques which focus on worst-case behavior.

Benchmarking Smart Meter Data Analytics

Smart electricity meters have been replacing conventional meters worldwide, enabling automated collection of fine-grained (every 15 minutes or hourly) consumption data. A variety of smart meter analytics algorithms and applications have been proposed, mainly in the smart grid literature, but the focus thus far has been on what can be done with the data rather than how to do it efficiently. In this paper, we examine smart meter analytics from a software performance perspective. First, we propose a performance benchmark that includes common data analysis tasks on smart meter data. Second, since obtaining large amounts of smart meter data is difficult due to privacy issues, we present an algorithm for generating large realistic data sets from a small seed of real data. Third, we implement the proposed benchmark using five representative platforms: a traditional numeric computing platform (Matlab), a relational DBMS with a built-in machine learning toolkit (PostgreSQL/MADLib), a main-memory column store (“System C”), and two distributed data processing platforms (Hive and Spark). We compare the five platforms in terms of application development effort and performance on a multi-core machine as well as a cluster of 16 commodity servers. We have made the proposed benchmark and data generator freely available online.

Resource Allocation in User-Controlled Circuit-Switched Optical Networks

The concept of a user-controlled circuit-switched optical network is gaining popularity in an effort to fulfill the insatiable data transport needs of the scientific community. We consider the resource allocation challenges that arise in such a network, based on prior experience in developing a user-controlled lightpath management system. In particular, we examine problems related to partitioning of lightpaths and construction of end-to-end lightpaths in support of large data transfers. Through a simulation study, we show that bandwidth fragmentation is an important issue in partitioning of SONET circuits, and we demonstrate an effective countermeasure. We also explore novel optimization criteria for routing of end-to-end lightpaths, and present problem formulations with efficient polynomial time solutions.

Making objects writable

We devise a technique for augmenting shared objects in the standard n-process shared memory model with a linearizable Write{} operation, using bounded space and optimal worst-case step complexity. We provide a transformation of any shared object SW supporting only sequential Write{} operations into an object

Eventually consistent: not what you were expecting?

W

Smart Meter Data Analytics: Systems, Algorithms, and Benchmarking

that supports concurrent Write{} operations. This transformation requires O(n2) SW objects and O(n2) O(log n)-bit registers, and each method (including Write{}) has, up to a constant additive term, the same time complexity as the corresponding method on object