No description available.

This thesis considers mobile computation offloading where task completion times are subject to hard deadline constraints. Hard deadlines are difficult to meet in conventional computation offloading due to the stochastic nature of the wireless channels involved. Rather than using binary offload decisions, we permit concurrent remote and local job execution when it is needed to ensure task completion deadlines. The thesis addresses this problem for homogeneous Markovian wireless channels. Two online energy-optimal computation offloading algorithms, OnOpt and MultiOpt, are proposed. OnOpt uploads the job to the server continuously and MultiOpt uploads the job in separate parts, each of which requires a separate offload initiation decision. The energy optimality of the algorithms is shown by constructing a time-dilated absorbing Markov process and applying dynamic programming. Closed form results are derived for general Markovian channels. The Gilbert-Elliott channel model is used to show how a particular Markov chain structure can be exploited to compute optimal offload initiation times more efficiently. The performance of the proposed algorithms is compared to three others, namely, Immediate Offloading, Channel Threshold, and Local Execution. Performance results show that the proposed algorithms can significantly improve mobile device energy consumption compared to the other approaches while guaranteeing hard task execution deadlines. / Thesis / Master of Applied Science (MASc)

Serverless computing is a new paradigm, and it is becoming rapidly popular in Cloud computing. Serverless computing has interesting, unique properties that the unit of deployment and execution is a serverless function. Moreover, it introduces the new economic model pay-as-you-go billing model. It provides a high economic benefit from highly elastic resource provisioning to the application. However, it also accompanies the new challenges for serverless computing: (1) start-up time latency problem from relatively short function execution time, (2) high-security risk from highly consolidated environment, and (3) memory efficiency problem from unpredictable function invocations. These problems not only degrade performance but also lowers the economic benefits of Cloud providers. In this work, we propose VM-level pre-JIT snapshot and develop Fireworks to solve the three main challenges without any compromises. The key idea behind the VM-level preJIT snapshot is to leverage pre-JITted serverless function codes to reduce both start-up time and execution time of the function and improve memory efficiency by sharing the pre-JITted codes. Also, Fireworks can provide high-level isolation by storing the pre-JITted codes to the snapshot of microVM's snapshot. Our evaluation shows that Fireworks outperforms the state-of-art serverless platforms by 20.6× and memory efficiency up to 7.3×. / Master of Science / Serverless computing is the most popular in cloud computing. Contrary to its name, developers write and run their code on servers managed by cloud providers. The number of servers, required CPU, memory are automatically adjusted in proportion to the incoming traffic. Also, the users only pay for what they use and the pay-as-you-go attracts attention as new infrastructure. Serverless computing continues to evolve and it is being done as research from business to academic. There are many efforts to reduce cold start, which is the delay in creating the necessary resources when a serverless program runs first. The serverless platforms prepare resources in advance or provide lighter cloud resources. However, this can waste resources or increase a security threat. In this work, we propose a fast, efficient, and safe serverless framework. We use Just-In-Time (JIT) compilation, which can improve the performance of the interpreter languages which are widely used in the serverless. We keep the JIT-generated machine code in the snapshot for reuse. Besides, the security is guaranteed by the VM-level snapshot. In addition, the snapshot can be shared, increasing memory efficiency. Through our implementation and evaluation, we have shown that Fireworks improve up to 20 times in terms of cold start performance and more than 7 times in memory efficiency than state-of-the-art serverless platforms. We believe our research has made a new way to use the JIT and the snapshot in the serverless computing.

Call number: LD2668 .T4 1978 E34 / Master of Science

Šiame darbe aprašomos technologijos naudojamos debesų kompiuterijos platformose. Pilnai išanalizuojama nemokama debesies platforma Eucalyptus. Bandoma sukurti internetinių puslapių talpinimo paslaugą debesyje (PaaS paslauga), kuria naudotis galėtų daug vartotojų. Taip pat sudaromas planas kaip galėtų atrodyti panašių paslaugų perkėlimas į debesies infrastruktūras. Išnagrinėjus, kokios programinės įrangos reikia tokiai paslaugai teikti, paruošti pavyzdiniai instaliaciniai skriptai, nubraižytos schemos kaip tokia paslauga galėtų veikti ir kokias funkcijas, bei kokią naudą gauna galutinis vartotojas naudodamas tokią paslaugą. Suprojektuota sistema, kuri automatiškai turi rūpintis tokios paslaugos valdymu, bei stebėjimu. Pateikti tokios automatizuotos sistemos kodo pavyzdžiai. / This document describes the technologies used in cloud computing platforms. Also this work completely analyze cloud open free platform Eucalyptus. On this platform trying to create a web page hosting service in the cloud as a PaaS service, which could be used of many users. Also work describes the plan/scheme as it might be possible to transfer similiar services to the cloud infrastructure. Examination of which software must be provided the following services, preparing model system installation scripts, either as a scheme for such a service can operate and what functions and what benefits the final consumer gets using this service. Designed a system that automatically can provide such a service management and monitoring. Shows such an automated system code examples.

Focus of this diploma thesis is on cloud computing topic. Cloud computing is current trend that changes the method of IT services providing. The concept of cloud computing model is based on a provision of IT resources through services. This model offers companies many advantages. The aim of this diploma thesis is to provide an overview of the motivational forces, barriers and their solution in the transition of small and medium businesses to products in the cloud. The theoretical part deals with economic and technological aspects arising from the use of cloud computing model. The next stage is provided with an overview of adoption of cloud computing services in the Czech Republic and current trends. The practical part is focused on the research of motivation aspects and barriers resulting from adoption services built on the cloud computing concept. The main contribution of this thesis is primarily in providing an overview of motivational aspects and research-based barriers in the environment of Czech companies. This overview could help to these company representatives, who are considering deployment of cloud based services, but are struggling with the barriers that this model brings.

Marine stratocumulus cloud regimes exert a strong climatic influence through their high solar reflectivity. Human-induced changes in stratocumulus clouds, attributed to an increase of the aerosol burden (indirect effects), can be significant given the cloud decks proximity to the continents; nevertheless, the magnitude and the final climatic consequences of these changes are uncertain. This thesis investigates further the interactions between aerosols, cloud microphysics, regional circulation, and radiative response in the Southeast Pacific stratocumulus cloud deck, one of the largest and most persistent cloud regimes in the planet. Specifically, three different aspects are addressed by this thesis: The importance of the synoptic atmospheric variability in controlling cloud microphysical and radiative changes, a validation analysis of satellite retrievals of cloud microphysics from MOderate Resolution Imaging Spectroradiometer (MODIS), and the quantitative assessments of cloud aerosol interactions along with their associated radiative forcing using primarily aircraft remote sensing data. Synoptic and satellite-derived cloud property variations for the Southeast Pacific region associated with changes in coastal satellite-derived cloud droplet number concentration (Nd) are analyzed through a composite technique. MAX and MIN Nd composites are defined by the top and bottom terciles of daily area-mean Nd values over the Arica Bight, the region with the largest mean oceanic Nd, for the five October months of 2001, 2005, 2006, 2007, and 2008. The MAX-Nd composite is characterized by a weaker subtropical anticyclone and weaker winds than the MIN-Nd composite. Additionally, the MAX-Nd composite clouds over the Arica Bight are thinner than the MIN-Nd composite clouds, have lower cloud tops, lower near-coastal cloud albedos, and occur below warmer and drier free tropospheres. At 85˚W, the top-of-atmosphere shortwave fluxes are significantly higher (50%) for the MAX-Nd, with thicker, lower clouds and higher cloud fractions than for the MIN-Nd. The change in Nd at this location is small, suggesting that the MAX-MIN Nd composite differences in radiative properties primarily reflects synoptic changes. The ability of MODIS level 2 retrievals to represent the cloud microphysics is assessed with in-situ measurements of droplet size distributions, collected during the VAMOS Ocean-Cloud-Atmosphere-Land Study Regional Experiment (VOCALS-REx). The MODIS cloud optical thickness (t) correlates well with the in-situ values with a positive bias (1.42). In contrast, the standard 2.1 micron-derived MODIS cloud effective radius (r_e) is found to systematically exceed the in-situ cloud-top r_e, with a mean bias of 2.08 um. Three sources of errors that could contribute to the MODIS r_e positive bias are investigated further: the spread of the cloud droplet size distribution, the presence of a separate drizzle mode, and the sensor viewing angles. The sensor zenith viewing angles were found to have little impact, while the algorithm assumption about the cloud droplet spectra and presence of a precipitation mode could affect the retrievals but not by enough to fully explain the positive MODIS r_e bias. The droplet spectra effects account for r_e offsets smaller than 0.6 um, 0.9 um, and 1.6 um for non-drizzling, light-drizzling, and heavy-drizzling clouds respectively. An explanation for the observed MODIS bias is lacking although three-dimensional radiative effects were not considered. This investigation supports earlier studies documenting a similar bias, this time using data from newer probes. MODIS r_e and t were also combined to estimate a liquid water path (LWP) and Nd. A positive bias was also apparent in LWP, and attributed to r_e. However, when selected appropriate parameters a priori, the MODIS Nd estimate was found to agree the best with the insitu aircraft observations of the four MODIS variables. Lastly, the first aerosol indirect effect (Twomey effect) is explicitly investigated with VOCALS-REx observations, collected during three daytime research flights (Nov 9, 11, and 13), utilizing an aerosol-cloud interactions metric, and defined as ACI=dln(t)/dln(Na), with Na corresponding to the accumulation mode aerosol concentration, t derived from a broadband pyranometer, and ACI binned by cloud LWP derived from a millimeter-wavelength radiometer. Aircraft remote sensing estimates of the ACI, during sub-cloud transects, show that the cloud aerosol-interactions are strong and close to the maximum theoretical value for thin clouds, with a decrease of ACI with LWP. Although an explanation for the dependence of ACI on LWP is lacking, we found that a decrease in ACI with LWP is associated with decreases in both surface meridional winds and Nd. Similar to ACI, albedo fractional changes due to Nd fractional changes also tended to be smaller for higher LWPs, but with an overall radiative forcing larger than conservative global estimates obtained in global circulation models. The findings of this thesis emphasize the strong stratocumulus albedo response to an aerosol perturbation and its dependence on the regional scale atmospheric configuration. The results presented here can be used as a benchmark for testing regional and climate models, as well as helping to improve the current parameterizations of the first aerosol indirect effect.

Context: Cloud Outsourcing is a new form of outsourcing which is not more under implementation and yet to be implemented. It is a form of outsourcing in which software organizations outsource the work to e-freelancers available throughout the world using cloud services via the Internet. Software organizations handovers the respective task to the cloud and from the cloud e-freelancers undertake the development of task and then return back the finished task to the cloud. Organizations recollect the finished task from the cloud and verify it and then pay to the e-freelancer. Objectives: The aim of this study is to identify the sequence of activities involved during the entire process of cloud outsourcing and to find out the risks which are likely to be occurred during the implementation of this process. To prioritize the elicitated risks according to their probability of occurrence, impact and cost required to mitigate the corresponding risk. Methods: Data is collected by literature review and then the data is synthesized. On the other side interviews with practitioners are conducted to know the activities involved and to find out the risks that are likely to be occurred during the implementation of cloud outsourcing. After this, a survey is conducted in order to prioritize the risks and a standard risk analysis is conducted to know the risks which are likely to be occurred. Literature review is done using four databases including the literature from the year 1990 to till date. Results: Totally we have identified 21 risks that are likely to be occurred and 8 activities so far. By performing risk analysis we have presented the risks, which should be considered first and relevant counter measures are suggested to overcome them. / mr.kommineni@me.com, mr.kommineni1@gmail.com phone no. +919963420123

This report describes a bachelor thesis performed at Xcerion for their subsidiary CloudMe. CloudMe has a service for cloud computing which makes your files accessible from anywhere, anytime. Cloud computing is addressed to both business and ordinary home users. With the ability to store important documents or family photos in the cloud other important subjects can be focused on. A cloud can in short be described as a metaphor for the internet and with the growing server houses and accessibility of cloud computing the need for USB-sticks (Universal Serial Bus) and hard drives are slowly fading away. CloudMe today offers several different services to access the cloud from a number of devices which includes smartphones, a web-browser and the home computer. This thesis work is about upgrading and improving their desktop application, Easy Upload. Easy Upload makes it possible to back-up several folders to the cloud but with the only option of uploading files and folders. Only uploading files to the cloud makes the application very restricted and lacking of an important feature, downloading from the cloud. Designing this new feature resulted in four different synchronization options; Upload, Download, Bidirectional and Hotsync, each configurable for every connected folder. With these options a user is offered the possibility to choose which synchronization best to use for the task in hand. The ability to choose, and configure, any folder on your computer opens up the restrictions which several large cloud computing services has today with one specific folder for the entire cloud. To ease the access and the understanding of these new synchronization options, a complete redesign of the user interface was developed. The name Easy Upload could no longer be associated with this application and therefore Easy Upload was renamed to suit the new functionality of the application and bring it closer to the association with the company CloudMe. When this thesis work was completed Easy Upload was considered to be the predecessor of CloudMe Sync.

Many enterprises are currently exploring the possibility of migrating some or all of their IT functionalities to public clouds with the objective of reducing their overall IT service costs or to open new business frontiers. Unfortunately, making such a decision is not a straightforward task; it requires a vigorous evaluation of the various benefits, risks and costs associated with the migration of their diverse business processes that comprise of their current IT services. Yet, this problem has received very little attention in the literature, mainly due to its interdisciplinary nature. This thesis aims at filling this gap by aiding the enterprises during the phase of making their cloud migration decision. The contributions of this work are twofold. First, a novel cloud-migration framework is introduced to guide the enterprises through a sequence of well-defined recommended analysis steps. These steps culminate with the formulation of the migration decision problem as a mathematical optimization one. The second contribution is a decision engine that efficiently solves this optimization problem. More precisely, the proposed framework gradually guides the enterprise to first identify the various business processes that are related to their IT services and then to determine the relationship and the communication needed among those processes. The identified inter-process communication represents an indicator of how tightly coupled these business processes are to each other. When outsourcing business processes, tightly coupled processes add a high communication cost and may introduce service latency if they are not co-located. As such, inter-process communication becomes an important input parameter that affects the migration decision. Enterprises can then determine to partially or completely migrate IT services to clouds. Furthermore, multiple vendors can be used for different services. However, when different vendors are involved, the communication cost between different processes increases. The objective is to maximize profit for an organization which includes lowering IT expenses in the long term without compromising data integrity or security. An optimization formula is finally constructed to help the enterprise determine which services to migrate given input parameters of the cost of doing business in-house, cost of outsourcing, and communication costs. Finally, a case study is utilized to demonstrate the performance of the proposed work by analyzing the process of migrating the services to clouds for an IPTV service provider. More specifically, the case study focuses on the content delivery network (CDN) within the IPTV provider’s infrastructure which is responsible for delivering contents to viewers. The CDN network can use the proposed profit-optimization formula to determine whether to utilize a cloud service or to use its internal resource to deliver the content. A performance evaluation from a simulation is presented to demonstrate the proposed profit-optimization formula can return a set of optimal mix of both internal and external services to maximize profits.