Ripduman Sohan

Predicting the Performance of Virtual Machine Migration

With the ability to move virtual machines between physical hosts, live migration is a core feature of virtualisation. However for migration to be useful, deployable feature on a large (datacentre) scale, we need to predict migration times with accuracy. In this paper, we characterise the parameters affecting live migration with particular emphasis on the Xen virtualisation platform. We discuss the relationships between the important parameters that affect migration and highlight how migration performance can vary considerably depending on the workload. We further provide 2 simulation models that are able to predict migration times to within 90% accuracy for both synthetic and real-world benchmarks.

An Analysis of Hard Drive Energy Consumption

The increasing storage capacity and necessary redundancy of data centers and other large-scale IT facilities has drawn attention to the issue of reducing the power consumption of hard drives. This work comprehensively investigates the power consumption of hard drives to determine typical runtime power profiles. We have instrumented at a fine-grained level and present our findings which show that (i) the energy consumed by the electronics of a drive is just as important as the mechanical energy consumption; (ii) the energy required to access data is affected by physical location on a drive; and (iii) the size of data transfers has measurable effect on power consumption.

A primer on provenance

Assessing the quality or validity of a piece of data is not usually done in isolation. You typically examine the context in which the data appears and try to determine its original sources or review the process through which it was created. This is not so straightforward when dealing with digital data, however: the result of a computation might have been derived from numerous sources and by applying complex successive transformations, possibly over long periods of time.

Melange: creating a "functional" internet

Most implementations of critical Internet protocols are written in type-unsafe languages such as C or C++ and are regularly vulnerable to serious security and reliability problems. Type-safe languages eliminate many errors but are not used to due to the perceived performance overheads. We combine two techniques to eliminate this performance penalty in a practical fashion: strong static typing and generative meta-programming. Static typing eliminates run-time type information by checking safety at compile-time and minimises dynamic checks. Meta-programming uses a single specification to abstract the low-level code required to transmit and receive packets. Our domain-specific language, MPL, describes Internet packet protocols and compiles into fast, zero-copy code for both parsing and creating these packets. MPL is designed for implementing quirky Internet protocols ranging from the low-level: Ethernet, IPv4, ICMP and TCP; to the complex application-level: SSH, DNS and BGP; and even file-system protocols such as 9P. We report on fully-featured SSH and DNS servers constructed using MPL and our OCaml framework Melange, and measure greater throughput, lower latency, better flexibility and more succinct source code than their C equivalents OpenSSH and BIND. Our quantitative analysis shows that the benefits of MPL-generated code overcomes the additional overheads of automatic garbage collection and dynamic bounds checking. Qualitatively, the flexibility of our approach shows that dramatic optimisations are easily possible.

Hard drive power consumption uncovered

Power consumption is a problem affecting all forms of computing, from server farms to mobile devices. Hard disks account for a significant percentage of a machines power consumption due to the mechanical nature of drive operation and increasingly sophisticated electronics. Due to this fact, there has been much research conducted with aims at reducing the power consumption of hard drives; examples including adaptive spin-down policies [1] and probabilistic management approaches [4]. However, this work has been done without fine-grained measurements of drive power consumption to accurately characterize trends; a shortcoming observed by other authors [3].

A user-level approach to network attached storage

Modern network attached storage architectures suffer from poor performance, resource management and inflexibility. We examine the problems of current architectures and propose an alternative, built around the precept of doing network storage at user-level. We further examine some of the challenges associated with this unusual approach and provide some preliminary block layer results indicating our architecture results in better performance than current designs

Intelligent Energy-Aware Networks

Today the energy consumption of Information and Communication Technology (ICT) industry is a significant contributor to the total energy demand in many developed countries. Recent studies show that the ICT industry is responsible for about 2% of the global emission of CO2 and this percentage is predicted to increase as the Internet expands in bandwidth and reach. In this chapter we highlight different approaches for energy efficiency in communication networks. Firstly we review the techniques proposed to reduce the energy consumption of communication networks at the equipment and network levels. Secondly we investigate the use of renewable energy to reduce the CO2 emission of IP over WDM networks. Issues including how to use renewable energy (solar in this work) more effectively, how to reduce the nonrenewable energy consumption of transponders (the second most energy consuming device in a node), how to select the location of nodes using renewable energy, and load dependent energy consumption are considered. Thirdly we discuss workload migration using virtualization technologies in data centers as an approach of energy consumption minimization. Finally we consider some of the photonic systems advances which have the potential to reduce significantly the energy consumption within Ethernet switches and IP routers in the datacenter, showing how integrated photonic switch fabrics are starting to have the performance required for energy efficient high switching applications.This two-volume work levels both criticism and challenge to traditional developmental psychology. For too long, developmental psychologists have been studying individuals as if they developed in a sociocultural vacuum. As psychologists began to study the individuals development more broadly, they considered the impact of a number of other factors in the physical and social environment: early education, sociocultural differences, mass communication, alternative living arrangements, and medical care-to name but a few. Volume I, Historical and Cultural Issues, examines the problems of behavioral development from historical, political, theoretical, and cultural points of view. A number of content areas already familiar to developmental psychologists are discussed: Piagets theory, perceptual development, socialization, and language acquisition. In addition, topics relatively unfamiliar to American psychologists are included: the contribution of early European developmentalists such as William and Clara Stern, Alfred Binet, and Eduard Spranger; and an introduction to recent Soviet developmental theory. Volume II, Social and Environmental Issues, considers the effects of changes in social and environmental conditions upon individual development. The expanding impact of technology such as the communications media, the importance of nutrition, and the design of playgrounds and other spaces for growing children are among the changes examined, as are the impact of social organizations and interactions within small groups, focusing upon preschool education, interaction within the family, and personality development throughout the individuals life.

Non-repudiable disk I/O in untrusted kernels

It is currently impossible for an application to verify that the data it passes to the kernel for storage is actually submitted to an underlying device or that the data returned to an application by the kernel has actually originated from an underlying device. A compromised or malicious OS can silently discard data written by the application or return fabricated data during a read operation. This is a serious data integrity issue for use-cases where verifiable storage and retrieval of data is a necessary precondition for ensuring correct operation, for example with secure logging, APT monitoring and compliance. We outline a solution for verifiable data storage and retrieval by providing a trustworthy mechanism, based on Intel SGX, to authenticate and verify request data at both the application and storage device endpoints. Even in the presence of a malicious OS our design ensures the authenticity and integrity of data while performing disk I/O and detects any data loss attributable to the untrusted OS fabricating or discarding read and write requests respectively. We provide a nascent prototype implementation for the core system together with an evaluation highlighting the temporal overheads imposed by this mechanism.

Applying Provenance in APT Monitoring and Analysis: Practical Challenges for Scalable, Efficient and Trustworthy Distributed Provenance

Advanced Persistent Threats (APT) are a class of security threats in which a well-resourced attacker targets a specific individual or organisation with a predefined goal. This typically involves exfiltration of confidential material, although increasingly attacks target the encryption or destruction of mission critical data. With traditional prevention and detection mechanisms failing to stem the tide of such attacks, there is a pressing need for new monitoring and analysis tools that reduce both false-positive rates and the cognitive burden on human analysts. We propose that local and distributed provenance metadata can simplify and improve monitoring and analysis of APTs by providing a single, authoritative sequence of events that captures the context (and side effects) of potentially malicious activities. Provenance metadata allows a human analyst to backtrack from detection of malicious activity to the point of intrusion and, similarly, to work forward to fully understand the consequences. Applying provenance to APT monitoring and analysis introduces some significantly different challenges and requirements in comparison to more traditional applications. Drawing from our experiences working with and adapting the OPUS (Observed Provenance in User Space) system to an APT monitoring and analysis use case, we introduce and discuss some of the key challenges in this space. These preliminary observations are intended Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. TaPP 2017, June 22-23, 2017, Seattle, Washington. Copyright remains with the owner/author(s). to prime a discussion within the community about the design space for scalable, efficient and trustworthy distributed provenance for scenarios that impose different constraints from traditional provenance applications such as workflow and data processing frameworks. CCS Concepts • Information systems → Data provenance; •Software and its engineering → Distributed systems organizing principles; •Security and privacy → Distributed systems security

Structural Analysis of Whole-System Provenance Graphs.

System based provenance generates traces captured from various systems, a representation method for inferring these traces is a graph. These graphs are not well understood, and current work focuses on their extraction and processing, without a thorough characterization being in place. This paper studies the topology of such graphs. We analyze multiple Whole-system-Provenance graphs and present that they have hubs-and-authorities model of graphs as well as a power law distribution. Our observations allow for a novel understanding of the structure of Whole-system-Provenance graphs.