Showing papers on "Server published in 2009"

Above the Clouds: A Berkeley View of Cloud Computing

Abstract: Cloud Computing, the long-held dream of computing as a utility, has the potential to transform a large part of the IT industry, making software even more attractive as a service and shaping the way IT hardware is designed and purchased. Developers with innovative ideas for new Internet services no longer require the large capital outlays in hardware to deploy their service or the human expense to operate it. They need not be concerned about overprovisioning for a service whose popularity does not meet their predictions, thus wasting costly resources, or underprovisioning for one that becomes wildly popular, thus missing potential customers and revenue. Moreover, companies with large batch-oriented tasks can get results as quickly as their programs can scale, since using 1000 servers for one hour costs no more than using one server for 1000 hours. This elasticity of resources, without paying a premium for large scale, is unprecedented in the history of IT. Cloud Computing refers to both the applications delivered as services over the Internet and the hardware and systems software in the datacenters that provide those services. The services themselves have long been referred to as Software as a Service (SaaS). The datacenter hardware and software is what we will call a Cloud. When a Cloud is made available in a pay-as-you-go manner to the general public, we call it a Public Cloud; the service being sold is Utility Computing. We use the term Private Cloud to refer to internal datacenters of a business or other organization, not made available to the general public. Thus, Cloud Computing is the sum of SaaS and Utility Computing, but does not include Private Clouds. People can be users or providers of SaaS, or users or providers of Utility Computing. We focus on SaaS Providers (Cloud Users) and Cloud Providers, which have received less attention than SaaS Users. From a hardware point of view, three aspects are new in Cloud Computing.

VL2: a scalable and flexible data center network

Abstract: To be agile and cost effective, data centers should allow dynamic resource allocation across large server pools. In particular, the data center network should enable any server to be assigned to any service. To meet these goals, we present VL2, a practical network architecture that scales to support huge data centers with uniform high capacity between servers, performance isolation between services, and Ethernet layer-2 semantics. VL2 uses (1) flat addressing to allow service instances to be placed anywhere in the network, (2) Valiant Load Balancing to spread traffic uniformly across network paths, and (3) end-system based address resolution to scale to large server pools, without introducing complexity to the network control plane. VL2's design is driven by detailed measurements of traffic and fault data from a large operational cloud service provider. VL2's implementation leverages proven network technologies, already available at low cost in high-speed hardware implementations, to build a scalable and reliable network architecture. As a result, VL2 networks can be deployed today, and we have built a working prototype. We evaluate the merits of the VL2 design using measurement, analysis, and experiments. Our VL2 prototype shuffles 2.7 TB of data among 75 servers in 395 seconds - sustaining a rate that is 94% of the maximum possible.

Globally Distributed Utility Computing Cloud

Abstract: Teachings of this application include a computing network that may include multiple different data centers and/or server grids which are deployed in different geographic locations. In at least one embodiment, at least some of the server grids may be operable to provide on-demand, grid and/or utility computing resources for hosting various types of distributed applications. In at least one embodiment, a distributed application may be characterized as an application made up of distinct components (e.g., virtual appliances, virtual machines, virtual interfaces, virtual volumes, virtual network connections, etc.) in separate runtime environments. In at least one embodiment, different ones of the distinct components of the distributed application may be hosted or deployed on different platforms (e.g., different servers) connected via a network. In some embodiments, a distributed application may be characterized as an application that runs on two or more networked computers.

BCube: a high performance, server-centric network architecture for modular data centers

Abstract: This paper presents BCube, a new network architecture specifically designed for shipping-container based, modular data centers. At the core of the BCube architecture is its server-centric network structure, where servers with multiple network ports connect to multiple layers of COTS (commodity off-the-shelf) mini-switches. Servers act as not only end hosts, but also relay nodes for each other. BCube supports various bandwidth-intensive applications by speeding-up one-to-one, one-to-several, and one-to-all traffic patterns, and by providing high network capacity for all-to-all traffic.BCube exhibits graceful performance degradation as the server and/or switch failure rate increases. This property is of special importance for shipping-container data centers, since once the container is sealed and operational, it becomes very difficult to repair or replace its components.Our implementation experiences show that BCube can be seamlessly integrated with the TCP/IP protocol stack and BCube packet forwarding can be efficiently implemented in both hardware and software. Experiments in our testbed demonstrate that BCube is fault tolerant and load balancing and it significantly accelerates representative bandwidth-intensive applications.

The nature of data center traffic: measurements & analysis

Abstract: We explore the nature of traffic in data centers, designed to support the mining of massive data sets. We instrument the servers to collect socket-level logs, with negligible performance impact. In a 1500 server operational cluster, we thus amass roughly a petabyte of measurements over two months, from which we obtain and report detailed views of traffic and congestion conditions and patterns. We further consider whether traffic matrices in the cluster might be obtained instead via tomographic inference from coarser-grained counter data.

Power and performance management of virtualized computing environments via lookahead control

Abstract: There is growing incentive to reduce the power consumed by large-scale data centers that host online services such as banking, retail commerce, and gaming. Virtualization is a promising approach to consolidating multiple online services onto a smaller number of computing resources. A virtualized server environment allows computing resources to be shared among multiple performance-isolated platforms called virtual machines. By dynamically provisioning virtual machines, consolidating the workload, and turning servers on and off as needed, data center operators can maintain the desired quality-of-service (QoS) while achieving higher server utilization and energy efficiency. We implement and validate a dynamic resource provisioning framework for virtualized server environments wherein the provisioning problem is posed as one of sequential optimization under uncertainty and solved using a lookahead control scheme. The proposed approach accounts for the switching costs incurred while provisioning virtual machines and explicitly encodes the corresponding risk in the optimization problem. Experiments using the Trade6 enterprise application show that a server cluster managed by the controller conserves, on average, 22% of the power required by a system without dynamic control while still maintaining QoS goals. Finally, we use trace-based simulations to analyze controller performance on server clusters larger than our testbed, and show how concepts from approximation theory can be used to further reduce the computational burden of controlling large systems.

Dynamic provable data possession

Abstract: We consider the problem of efficiently proving the integrity of data stored at untrusted servers. In the provable data possession (PDP) model, the client preprocesses the data and then sends it to an untrusted server for storage, while keeping a small amount of meta-data. The client later asks the server to prove that the stored data has not been tampered with or deleted (without downloading the actual data). However, the original PDP scheme applies only to static (or append-only) files.We present a definitional framework and efficient constructions for dynamic provable data possession (DPDP), which extends the PDP model to support provable updates to stored data. We use a new version of authenticated dictionaries based on rank information. The price of dynamic updates is a performance change from O(1) to O(logn) (or O(nelog n), for a file consisting of n blocks, while maintaining the same (or better, respectively) probability of misbehavior detection. Our experiments show that this slowdown is very low in practice (e.g. 415KB proof size and 30ms computational overhead for a 1GB file). We also show how to apply our DPDP scheme to outsourced file systems and version control systems (e.g. CVS).

Ensuring data storage security in Cloud Computing

Abstract: Cloud Computing has been envisioned as the next-generation architecture of IT Enterprise. In contrast to traditional solutions, where the IT services are under proper physical, logical and personnel controls, Cloud Computing moves the application software and databases to the large data centers, where the management of the data and services may not be fully trustworthy. This unique attribute, however, poses many new security challenges which have not been well understood. In this article, we focus on cloud data storage security, which has always been an important aspect of quality of service. To ensure the correctness of users' data in the cloud, we propose an effective and flexible distributed scheme with two salient features, opposing to its predecessors. By utilizing the homomorphic token with distributed verification of erasure-coded data, our scheme achieves the integration of storage correctness insurance and data error localization, i.e., the identification of misbehaving server(s). Unlike most prior works, the new scheme further supports secure and efficient dynamic operations on data blocks, including: data update, delete and append. Extensive security and performance analysis shows that the proposed scheme is highly efficient and resilient against Byzantine failure, malicious data modification attack, and even server colluding attacks.

HAIL: a high-availability and integrity layer for cloud storage

Abstract: We introduce HAIL (High-Availability and Integrity Layer), a distributed cryptographic system that allows a set of servers to prove to a client that a stored file is intact and retrievable. HAIL strengthens, formally unifies, and streamlines distinct approaches from the cryptographic and distributed-systems communities. Proofs in HAIL are efficiently computable by servers and highly compact---typically tens or hundreds of bytes, irrespective of file size. HAIL cryptographically verifies and reactively reallocates file shares. It is robust against an active, mobile adversary, i.e., one that may progressively corrupt the full set of servers. We propose a strong, formal adversarial model for HAIL, and rigorous analysis and parameter choices. We show how HAIL improves on the security and efficiency of existing tools, like Proofs of Retrievability (PORs) deployed on individual servers. We also report on a prototype implementation.

RouteBricks: exploiting parallelism to scale software routers

Abstract: We revisit the problem of scaling software routers, motivated by recent advances in server technology that enable high-speed parallel processing--a feature router workloads appear ideally suited to exploit. We propose a software router architecture that parallelizes router functionality both across multiple servers and across multiple cores within a single server. By carefully exploiting parallelism at every opportunity, we demonstrate a 35Gbps parallel router prototype; this router capacity can be linearly scaled through the use of additional servers. Our prototype router is fully programmable using the familiar Click/Linux environment and is built entirely from off-the-shelf, general-purpose server hardware.

Optimal power allocation in server farms

Abstract: Server farms today consume more than 1.5% of the total electricity in the U.S. at a cost of nearly $4.5 billion. Given the rising cost of energy, many industries are now seeking solutions for how to best make use of their available power. An important question which arises in this context is how to distribute available power among servers in a server farm so as to get maximum performance.By giving more power to a server, one can get higher server frequency (speed). Hence it is commonly believed that, for a given power budget, performance can be maximized by operating servers at their highest power levels. However, it is also conceivable that one might prefer to run servers at their lowest power levels, which allows more servers to be turned on for a given power budget. To fully understand the effect of power allocation on performance in a server farm with a fixed power budget, we introduce a queueing theoretic model, which allows us to predict the optimal power allocation in a variety of scenarios. Results are verified via extensive experiments on an IBM BladeCenter.We find that the optimal power allocation varies for different scenarios. In particular, it is not always optimal to run servers at their maximum power levels. There are scenarios where it might be optimal to run servers at their lowest power levels or at some intermediate power levels. Our analysis shows that the optimal power allocation is non-obvious and depends on many factors such as the power-to-frequency relationship in the processors, the arrival rate of jobs, the maximum server frequency, the lowest attainable server frequency and the server farm configuration. Furthermore, our theoretical model allows us to explore more general settings than we can implement, including arbitrarily large server farms and different power-to-frequency curves. Importantly, we show that the optimal power allocation can significantly improve server farm performance, by a factor of typically 1.4 and as much as a factor of 5 in some cases.

Fast and accurate automatic structure prediction with HHpred

Abstract: Automated protein structure prediction is becoming a mainstream tool for biological research. This has been fueled by steady improvements of publicly available automated servers over the last decade, in particular their ability to build good homology models for an increasing number of targets by reliably detecting and aligning more and more remotely homologous templates. Here, we describe the three fully automated versions of the HHpred server that participated in the community-wide blind protein structure prediction competition CASP8. What makes HHpred unique is the combination of usability, short response times (typically under 15 min) and a model accuracy that is competitive with those of the best servers in CASP8.

Hyrax: Cloud Computing on Mobile Devices using MapReduce

Abstract: : Today's smartphones operate independently of each other, using only local computing, sensing, networking, and storage capabilities and functions provided by remote Internet services. It is generally difficult or expensive for one smartphone to share data and computing resources with another. Data is shared through centralized services, requiring expensive uploads and downloads that strain wireless data networks. Collaborative computing is only achieved using ad hoc approaches. Coordinating smartphone data and computing would allow mobile applications to utilize the capabilities of an entire smartphone cloud while avoiding global network bottlenecks. In many cases, processing mobile data in-place and transferring it directly between smartphones would be more efficient and less susceptible to network limitations than off loading data and processing to remote servers. We have developed Hyrax, a platform derived from Hadoop that supports cloud computing on Android smartphones. Hyrax allows client applications to conveniently utilize data and execute computing jobs on networks of smartphones and heterogeneous networks of phones and servers. By scaling with the number of devices and tolerating node departure, Hyrax allows applications to use distributed resources abstractly, oblivious to the physical nature of the cloud. The design and implementation of Hyrax is described, including experiences in porting Hadoop to the Android platform and the design of mobile specific customizations. The scalability of Hyrax is evaluated experimentally and compared to that of Hadoop. Although the performance of Hyrax is poor for CPU-bound tasks, it is shown to tolerate node-departure and offer reasonable performance in data sharing. A distributed multimedia search and sharing application is implemented to qualitatively evaluate Hyrax from an application development perspective.

Universal Blind Quantum Computation

Abstract: We present a protocol which allows a client to have a server carry out a quantum computation for her such that the client's inputs, outputs and computation remain perfectly private, and where she does not require any quantum computational power or memory. The client only needs to be able to prepare single qubits randomly chosen from a finite set and send them to the server, who has the balance of the required quantum computational resources. Our protocol is interactive: after the initial preparation of quantum states, the client and server use two-way classical communication which enables the client to drive the computation, giving single-qubit measurement instructions to the server, depending on previous measurement outcomes. Our protocol works for inputs and outputs that are either classical or quantum. We give an authentication protocol that allows the client to detect an interfering server; our scheme can also be made fault-tolerant. We also generalize our result to the setting of a purely classical client who communicates classically with two non-communicating entangled servers, in order to perform a blind quantum computation. By incorporating the authentication protocol, we show that any problem in BQP has an entangled two-prover interactive proof with a purely classical verifier. Our protocol is the first universal scheme which detects a cheating server, as well as the first protocol which does not require any quantum computation whatsoever on the client's side. The novelty of our approach is in using the unique features of measurement-based quantum computing which allows us to clearly distinguish between the quantum and classical aspects of a quantum computation.

Data center interconnect and traffic engineering

Abstract: A system for commoditizing data center networking is disclosed. The system includes an interconnection topology for a data center having a plurality of servers and a plurality of nodes of a network in the data center through which data packets may be routed. The system uses a routing scheme where the routing is oblivious to the traffic pattern between nodes in the network, and wherein the interconnection topology contains a plurality of paths between one or more servers. The multipath routing may be Valiant load balancing. It disaggregates the function of load balancing into a group of regular servers, with the result that load balancing server hardware can be distributed amongst racks in the data center leading to greater agility and less fragmentation. The architecture creates a huge, flexible switching domain, supporting any server/any service, full mesh agility, and unregimented server capacity at low cost.

Applying NOX to the Datacenter.

Abstract: Internet datacenters offer unprecedented computing power for a new generation of data-intensive computational tasks. There is a rapidly growing literature on the operating and distributed systems issues raised by these datacenters, but only recently have researchers turned their attention to the datacenter’s unique set of networking challenges. In contrast to enterprise networks, which usually grow organically over time, datacenter networks are carefully and coherently architected, so they provide an isolated setting in which new networking designs can be explored and deployed. The combination of interesting intellectual challenges and lowered deployment barriers makes datacenters a rich arena for network research and innovation, as evinced by the recent flurry of research papers on datacenter networks. Of particular interest are the network designs proposed in [1, 5, 6, 9], which vary along many design dimensions but are all specifically tailored to the datacenter environment. In the more general networking literature, in 2004 the 4D project [4] initiated a renaissance in the network management literature by advocating a logically centralized view of the network. The goal of this approach was to provide a general management plane, not specialized to a particular context (such as the datacenter). A recent development in this vein is the NOX network operating system [7]. NOX gives logically centralized access to high-level network abstractions such as users, topology, and services, and exerts control over the network by installing flow entries in switch forwarding tables. By providing programmatic access (through Python or C++) to network observation and control primitives, NOX serves as a flexible and scalable platform for building advanced network management functionality. Enterprise network management systems built on NOX have been in production use for over a year, and an early version of NOX is freely available under the GPL license at www.noxrepo.org. This philosophical question behind this paper is whether the general-purpose approach in networking, which has served the Internet and enterprise so well, can be extended to specialized environments like the datacenter, or if special-case solutions will prevail. The more practical instantiation of this question is: How well does a general-purpose management system, like NOX, cope with the highly specific and stringent requirements of the datacenter? As we explain in this paper, we find that not only can NOX provide reasonable management of datacenter environments, it also offers operators a choice of several points in the datacenter design spectrum, rather than locking them into one specific solution. Due to our familiarity with it, we use NOX throughout the paper as a concrete instance of a general network platform. However, the goal is to explore more broadly whether a general approach can be used in place of point solutions. Hence, this discussion should apply equally well to similar systems such as 4D [4], or Maestro [2]. We also don’t make the claim that these systems are better than any of these point solutions. Our only goal is to demonstrate that there is still hope for the “generalpurpose” philosophy that has served networking so well. In the rest of the paper, we first present background material on datacenter networks and NOX (§2), and then demonstrate NOX’s flexibility by describing implementations of existing architectures that can scale to a hundred thousand servers and millions of VMs (§3). We subsequently discuss how NOX provides basic datacenter functionality (§4) and several additional capabilities (§5). We end with a general discussion (§6).

Factored operating systems (fos): the case for a scalable operating system for multicores

Abstract: The next decade will afford us computer chips with 100's to 1,000's of cores on a single piece of silicon. Contemporary operating systems have been designed to operate on a single core or small number of cores and hence are not well suited to manage and provide operating system services at such large scale. If multicore trends continue, the number of cores that an operating system will be managing will continue to double every 18 months. The traditional evolutionary approach of redesigning OS subsystems when there is insufficient parallelism will cease to work because the rate of increasing parallelism will far outpace the rate at which OS designers will be capable of redesigning subsystems. The fundamental design of operating systems and operating system data structures must be rethought to put scalability as the prime design constraint. This work begins by documenting the scalability problems of contemporary operating systems. These studies are used to motivate the design of a factored operating system (fos). fos is a new operating system targeting manycore systems with scalability as the primary design constraint, where space sharing replaces time sharing to increase scalability.We describe fos, which is built in a message passing manner, out of a collection of Internet inspired services. Each operating system service is factored into a set of communicating servers which in aggregate implement a system service. These servers are designed much in the way that distributed Internet services are designed, but instead of providing high level Internet services, these servers provide traditional kernel services and replace traditional kernel data structures in a factored, spatially distributed manner. fos replaces time sharing with space sharing. In other words, fos's servers are bound to distinct processing cores and by doing so do not fight with end user applications for implicit resources such as TLBs and caches. We describe how fos's design is well suited to attack the scalability challenge of future multicores and discuss how traditional application-operating systems interfaces can be redesigned to improve scalability.

EnaCloud: An Energy-Saving Application Live Placement Approach for Cloud Computing Environments

Abstract: With the increasing prevalence of large scale cloud computing environments, how to place requested applications into available computing servers regarding to energy consumption has become an essential research problem, but existing application placement approaches are still not effective for live applications with dynamic characters. In this paper, we proposed a novel approach named EnaCloud, which enables application live placement dynamically with consideration of energy efficiency in a cloud platform. In EnaCloud, we use a Virtual Machine to encapsulate the application, which supports applications scheduling and live migration to minimize the number of running machines, so as to save energy. Specially, the application placement is abstracted as a bin packing problem, and an energy-aware heuristic algorithm is proposed to get an appropriate solution. In addition, an over-provision approach is presented to deal with the varying resource demands of applications. Our approach has been successfully implemented as useful components and fundamental services in the iVIC platform. Finally, we evaluate our approach by comprehensive experiments based on virtual machine monitor Xen and the results show that it is feasible.

Dynamic Scaling of Web Applications in a Virtualized Cloud Computing Environment

Abstract: Scalability is critical to the success of many enterprises currently involved in doing business on the web and in providing information that may vary drastically from one time to another. Maintaining sufficient resources just to meet peak requirements can be costly. Cloud computing provides a powerful computing model that allows users to access resources on-demand. In this paper, we will describe a novel architecture for the dynamic scaling of web applications based on thresholds in a virtualized Cloud Computing environment. We will illustrate our scaling approach with a front-end load-balancer for routing and balancing user requests to web applications deployed on web servers installed in virtual machine instances. A dynamic scaling algorithm for automated provisioning of virtual machine resources based on threshold number of active sessions will be introduced. The on-demand capability of the Cloud to rapidly provision and dynamically allocate resources to users will be discussed. Our work has demonstrated the compelling benefits of the Cloud which is capable of handling sudden load surges, delivering IT resources on-demands to users, and maintaining higher resource utilization, thus reducing infrastructure and management costs.

Global server load balancing

Abstract: A global server load balancing (GSLB) switch serves as a proxy to an authoritative DNS communicates with numerous site switches which are coupled to host servers serving specific applications. The GSLB switch receives from site switches operational information regarding host servers within the site switches neighborhood. When a client program requests a resolution of a host name, the GSLB switch, acting as a proxy of an authoritative DNS, returns one or more ordered IP addresses for the host name. The IP addresses are ordered using metrics that include the information collected from the site switches. In one instance, the GSLB switch places the address that is deemed “best” at the top of the list.

Mobyle: a new full web bioinformatics framework

Abstract: Motivation: For the biologist, running bioinformatics analyses involves a time-consuming management of data and tools. Users need support to organize their work, retrieve parameters and reproduce their analyses. They also need to be able to combine their analytic tools using a safe data flow software mechanism. Finally, given that scientific tools can be difficult to install, it is particularly helpful for biologists to be able to use these tools through a web user interface. However, providing a web interface for a set of tools raises the problem that a single web portal cannot offer all the existing and possible services: it is the user, again, who has to cope with data copy among a number of different services. A framework enabling portal administrators to build a network of cooperating services would therefore clearly be beneficial. Results: We have designed a system, Mobyle, to provide a flexible and usable Web environment for defining and running bioinformatics analyses. It embeds simple yet powerful data management features that allow the user to reproduce analyses and to combine tools using a hierarchical typing system. Mobyle offers invocation of services distributed over remote Mobyle servers, thus enabling a federated network of curated bioinformatics portals without the user having to learn complex concepts or to install sophisticated software. While being focused on the end user, the Mobyle system also addresses the need, for the bioinfomatician, to automate remote services execution: PlayMOBY is a companion tool that automates the publication of BioMOBY web services, using Mobyle program definitions. Availability: The Mobyle system is distributed under the terms of the GNU GPLv2 on the project web site (http://bioweb2.pasteur.fr/projects/mobyle/). It is already deployed on three servers: http://mobyle.pasteur.fr, http://mobyle.rpbs.univ-paris-diderot.fr and http://lipm-bioinfo.toulouse.inra.fr/Mobyle. The PlayMOBY companion is distributed under the terms of the CeCILL license, and is available at http://lipm-bioinfo.toulouse.inra.fr/biomoby/PlayMOBY/. Contact: rf.ruetsap@troppus-elybom; rf.toredid-sirap-vinu.sbpr@troppus-elybom; rf.ruetsap@ladnotel Supplementary information:Supplementary data are available at Bioinformatics online.

A Theory of QoS for Wireless

Abstract: Wireless networks are increasingly used to carry applications with QoS constraints. Two problems arise when dealing with traffic with QoS constraints. One is admission control, which consists of determining whether it is possible to fulfill the demands of a set of clients. The other is finding an optimal scheduling policy to meet the demands of all clients. In this paper, we propose a framework for jointly addressing three QoS criteria: delay, delivery ratio, and channel reliability. We analytically prove the necessary and sufficient condition for a set of clients to be feasible with respect to the above three criteria. We then establish an efficient algorithm for admission control to decide whether a set of clients is feasible. We further propose two scheduling policies and prove that they are feasibility optimal in the sense that they can meet the demands of every feasible set of clients. In addition, we show that these policies are easily implementable on the IEEE 802.11 mechanisms. We also present the results of simulation studies that appear to confirm the theoretical studies and suggest that the proposed policies outperform others tested under a variety of settings.

Power-aware scheduling of virtual machines in DVFS-enabled clusters

Abstract: With the advent of Cloud computing, large-scale virtualized compute and data centers are becoming common in the computing industry. These distributed systems leverage commodity server hardware in mass quantity, similar in theory to many of the fastest Supercomputers in existence today. However these systems can consume a cities worth of power just to run idle, and require equally massive cooling systems to keep the servers within normal operating temperatures. This produces CO 2 emissions and significantly contributes to the growing environmental issue of Global Warming. Green computing, a new trend for high-end computing, attempts to alleviate this problem by delivering both high performance and reduced power consumption, effectively maximizing total system efficiency. This paper focuses on scheduling virtual machines in a compute cluster to reduce power consumption via the technique of Dynamic Voltage Frequency Scaling (DVFS). Specifically, we present the design and implementation of an efficient scheduling algorithm to allocate virtual machines in a DVFS-enabled cluster by dynamically scaling the supplied voltages. The algorithm is studied via simulation and implementation in a multi-core cluster. Test results and performance discussion justify the design and implementation of the scheduling algorithm.

Network consolidation for virtualized servers

Abstract: Techniques are disclosed for virtualized server kernel and virtual networks consolidation. The network consolidation allows a data center to migrate from an infrastructure that uses multiple dedicated gigabit Ethernet Network Adapters to manage system virtualization and migration to an infrastructure using consolidated, redundant, 10 gigabit Ethernet adapters. Different priority classes may be defined for different classes of network traffic such as hypervisor management traffic, inter-host virtual machine migration traffic, virtual machine production traffic, virtualized switching control plane traffic, etc. Further, an enhanced transmission standard may be used to specify a minimum bandwidth guarantee for certain traffic classes. Thus, the hypervisor management and inter-host virtual machine migration traffic may be transmitted, even the presence of congestion.

SLA-Aware Virtual Resource Management for Cloud Infrastructures

Abstract: Cloud platforms host several independent applications on a shared resource pool with the ability to allocate computing power to applications on a per-demand basis. The use of server virtualization techniques for such platforms provide great flexibility with the ability to consolidate several virtual machines on the same physical server, to resize a virtual machine capacity and to migrate virtual machine across physical servers. A key challenge for cloud providers is to automate the management of virtual servers while taking into account both high-level QoS requirements of hosted applications and resource management costs. This paper proposes an autonomic resource manager to control the virtualized environment which decouples the provisioning of resources from the dynamic placement of virtual machines. This manager aims to optimize a global utility function which integrates both the degree of SLA fulfillment and the operating costs. We resort to a Constraint Programming approach to formulate and solve the optimization problem. Results obtained through simulations validate our approach.

Shares and utilities based power consolidation in virtualized server environments

Abstract: Virtualization technologies like VMware and Xen provide features to specify the minimum and maximum amount of resources that can be allocated to a virtual machine (VM) and a shares based mechanism for the hypervisor to distribute spare resources among contending VMs. However much of the existing work on VM placement and power consolidation in data centers fails to take advantage of these features. One of our experiments on a real testbed shows that leveraging such features can improve the overall utility of the data center by 47% or even higher. Motivated by these, we present a novel suite of techniques for placement and power consolidation of VMs in data centers taking advantage of the min-max and shares features inherent in virtualization technologies. Our techniques provide a smooth mechanism for power-performance tradeoffs in modern data centers running heterogeneous applications, wherein the amount of resources allocated to a VM can be adjusted based on available resources, power costs, and application utilities. We evaluate our techniques on a range of large synthetic data center setups and a small real data center testbed comprising of VMware ESX servers. Our experiments confirm the end-to-end validity of our approach and demonstrate that our final candidate algorithm, PowerExpandMinMax, consistently yields the best overall utility across a broad spectrum of inputs - varying VM sizes and utilities, varying server capacities and varying power costs - thus providing a practical solution for administrators.

Drafting behind Akamai: inferring network conditions based on CDN redirections

Abstract: To enhance Web browsing experiences, content distribution networks (CDNs) move Web content "closer" to clients by caching copies of Web objects on thousands of servers worldwide. Additionally, to minimize client download times, such systems perform extensive network and server measurements and use them to redirect clients to different servers over short time scales. In this paper, we explore techniques for inferring and exploiting network measurements performed by the largest CDN, Akamai; our objective is to locate and utilize quality Internet paths without performing extensive path probing or monitoring. Our contributions are threefold. First, we conduct a broad measurement study of Akamai's CDN. We probe Akamai's network from 140 PlanetLab (PL) vantage points for two months. We find that Akamai redirection times, while slightly higher than advertised, are sufficiently low to be useful for network control. Second, we empirically show that Akamai redirections overwhelmingly correlate with network latencies on the paths between clients and the Akamai servers. Finally, we illustrate how large-scale overlay networks can exploit Akamai redirections to identify the best detouring nodes for one-hop source routing. Our research shows that in more than 50% of investigated scenarios, it is better to route through the nodes "recommended" by Akamai than to use the direct paths. Because this is not the case for the rest of the scenarios, we develop low-overhead pruning algorithms that avoid Akamai-driven paths when they are not beneficial. Because these Akamai nodes are part of a closed system, we provide a method for mapping Akamai-recommended paths to those in a generic overlay and demonstrate that these one-hop paths indeed outperform direct ones.

Autonomic virtual resource management for service hosting platforms

Abstract: Cloud platforms host several independent applications on a shared resource pool with the ability to allocate computing power to applications on a per-demand basis. The use of server virtualization techniques for such platforms provide great flexibility with the ability to consolidate several virtual machines on the same physical server, to resize a virtual machine capacity and to migrate virtual machine across physical servers. A key challenge for cloud providers is to automate the management of virtual servers while taking into account both high-level QoS requirements of hosted applications and resource management costs. This paper proposes an autonomic resource manager to control the virtualized environment which decouples the provisioning of resources from the dynamic placement of virtual machines. This manager aims to optimize a global utility function which integrates both the degree of SLA fulfillment and the operating costs. We resort to a Constraint Programming approach to formulate and solve the optimization problem. Results obtained through simulations validate our approach.

Content delivery for client server protocols with user affinities using connection end-point proxies

Abstract: In a network supporting transactions between clients and servers over a network path having operating characteristics to overcome, data is transported to overcome the operating characteristics using user affinities and dynamic user location information to selectively preload data, or representations, signatures, segments, etc. of data, in order to overcome the one or more operating characteristic. Examples of operating characteristics to overcome include bandwidth limitations, errors and latency. The dynamic location information can be stored in data structures accessible by agents of a data server and the data structures are populated based on user activities with respect to proxies associated with user locations, or the dynamic location information can be obtained implicitly as proxies maintain connections after termination by clients and the use of those maintained connections for preloading of data for the users associated with those clients. The data being preloaded can be protocol-specific data or protocol-independent data.

The effectiveness of deduplication on virtual machine disk images

Abstract: Virtualization is becoming widely deployed in servers to efficiently provide many logically separate execution environments while reducing the need for physical servers. While this approach saves physical CPU resources, it still consumes large amounts of storage because each virtual machine (VM) instance requires its own multi-gigabyte disk image. Moreover, existing systems do not support ad hoc block sharing between disk images, instead relying on techniques such as overlays to build multiple VMs from a single "base" image.Instead, we propose the use of deduplication to both reduce the total storage required for VM disk images and increase the ability of VMs to share disk blocks. To test the effectiveness of deduplication, we conducted extensive evaluations on different sets of virtual machine disk images with different chunking strategies. Our experiments found that the amount of stored data grows very slowly after the first few virtual disk images if only the locale or software configuration is changed, with the rate of compression suffering when different versions of an operating system or different operating systems are included. We also show that fixed-length chunks work well, achieving nearly the same compression rate as variable-length chunks. Finally, we show that simply identifying zero-filled blocks, even in ready-to-use virtual machine disk images available online, can provide significant savings in storage.