Muhammad Abdullah Adnan

Energy Efficient Geographical Load Balancing via Dynamic Deferral of Workload

With the increasing popularity of Cloud computing and Mobile computing, individuals, enterprises and research centers have started outsourcing their IT and computational needs to on-demand cloud services. Recently geographical load balancing techniques have been suggested for data centers hosting cloud computation in order to reduce energy cost by exploiting the electricity price differences across regions. However, these algorithms do not draw distinction among diverse requirements for responsiveness across various workloads. In this paper, we use the flexibility from the Service Level Agreements (SLAs) to differentiate among workloads under bounded latency requirements and propose a novel approach for cost savings for geographical load balancing. We investigate how much workload to be executed in each data center and how much workload to be delayed and migrated to other data centers for energy saving while meeting deadlines. We present an offline formulation for geographical load balancing problem with dynamic deferral and give online algorithms to determine the assignment of workload to the data centers and the migration of workload between data centers in order to adapt with dynamic electricity price changes. We compare our algorithms with the greedy approach and show that significant cost savings can be achieved by migration of workload and dynamic deferral with future electricity price prediction. We validate our algorithms on MapReduce traces and show that geographic load balancing with dynamic deferral can provide 20-30% cost-savings.

Path Consolidation for Dynamic Right-Sizing of Data Center Networks

Data center topologies typically consist of multirooted trees with many equal-cost paths between a given pair of hosts. Existing power optimization techniques do not utilize this property of data center networks for power proportionality. In this paper, we exploit this opportunity and show that significant energy savings can be achieved via path consolidation in the network. We present an offline formulation for the flow assignment in a data center network and develop an online algorithm by path consolidation for dynamic right-sizing of the network to save energy. To validate our algorithm, we build a flow level simulator for a data center network. Our simulation on flow traces generated from MapReduce workload shows 80% reduction in network energy consumption in data center networks and ~25% more energy savings compared to the existing techniques for saving energy in data center networks.

Minimum Segment Drawings of Series-Parallel Graphs with the Maximum Degree Three

A minimum segment drawing Γ of a planar graph G is a straight line drawing of G that has the minimum number of segments among all straight line drawings of G . In this paper, we give a linear-time algorithm for computing a minimum segment drawing of a series-parallel graph with the maximum degree three. To the best of our knowledge, this is the first algorithm for computing minimum segment drawings of an important subclass of planar graphs.

Utility-aware deferred load balancing in the cloud driven by dynamic pricing of electricity

Distributed computing resources in a cloud computing environment provides an opportunity to reduce energy and its cost by shifting loads in response to dynamically varying availability of energy. This variation in electrical power availability is represented in its dynamically changing price that can be used to drive workload deferral against performance requirements. But such deferral may cause user dissatisfaction. In this paper, we quantify the impact of deferral on user satisfaction and utilize flexibility from the service level agreements (SLAs) for deferral to adapt with dynamic price variation. We differentiate among the jobs based on their requirements for responsiveness and schedule them for energy saving while meeting deadlines and user satisfaction. Representing utility as decaying functions along with workload deferral, we make a balance between loss of user satisfaction and energy efficiency. We model delay as decaying functions and guarantee that no job violates the maximum deadline, and we minimize the overall energy cost. Our simulation on MapReduce traces show that energy consumption can be reduced by ∼15%, with such utility-aware deferred load balancing. We also found that considering utility as a decaying function gives better cost reduction than load balancing with a fixed deadline.

Two step verification system of highly secure social media: Possible to breach the security

With the increasing demand of social media, the security threats of these networks also increasing dramatically. Facebook having around 1.5 billion users are concerned with the security and privacy of its users. To make it more secure, the Facebook authority has introduced two-step verification system for login in ones account. However, the second step of verification, i.e., verification code based security, can easily be breached with a simple trick. In this paper, we will show how we breached Facebook security and hacked accounts within 6 to 10 seconds only and we will propose a solution regarding this limitation.

Distribution of distinguishable objects to bins: generating all distributions

In this paper we give an algorithm to generate all distributions of distinguishable objects to bins without repetition. Our algorithm generates each distribution in constant time. To the best of our knowledge, our algorithm is the first algorithm which generates each solution in O(1) time in the ordinary sense. As a byproduct of our algorithm, we obtain a new algorithm to enumerate all multiset partitions when the number of partitions is fixed and the partitions are numbered. In this case, the algorithm generates each multiset partitions in constant time (in the ordinary sense). Finally, we extend the algorithm to the case when the bins have priorities associated with them. Overall space complexity of the algorithm is O(mklgn), where there are m bins and the objects fall into k different classes. In a companion paper, the generation of all distributions of identical objects to bins is also considered.

sSketch: A Scalable Sketching Technique for PCA in the Cloud

Multidimensional data appear frequently in many web-related applications, e.g., product ratings, the bag-of-words representation of web pages, etc. Principal Component Analysis (PCA) has been widely used for discovering patterns in relationships among entities in multidimensional data. However, existing algorithms for PCA have limited scalability since they explicitly materialize intermediate data, whose size rapidly grows as the dimension increases. To avoid scalability issues, we propose sSketch, a scalable sketching technique for PCA that employs several optimization ideas, such as mean propagation, efficient sparse matrix operations, and effective job consolidation to minimize intermediate data. Using sSketch, we also provide two other scalable methods for deriving singular value and 2-norm of reconstruction error, both of which are used for data analysis purpose. We provide our implementation on popular Spark framework for distributed platform. We compare our method against state-of-the-art library functions available for distributed settings, namely MLlib-PCA and Mahout-PCA with real big datasets. Our experiments show that our method outperforms both of them by a wide margin. To encourage reproducibility, the source code of sSketch is made publicly available at \hrefhttps://github.com/DataMiningResearch/sSketch https://github.com/DataMiningResearch/sSketch.

Cooling energy integration in SimGrid

Cooling energy consumption is one of the most significant parts of total energy consumed by distributed systems. However, little effort has been spent so far to integrate the cooling energy in simulators that are used for simulating distributed systems. Therefore, in this paper, we propose an integration of cooling energy consumption in a widely-known simulator of distributed systems namely SimGrid. Here, we present necessary energy models that are needed to measure cooling energy required for a distributed system. Subsequently, we perform necessary modifications in SimGrid to integrate the models. We perform rigorous simulation over different settings using our integrated modules of SimGrid. Alongside, we perform real experimentation using settings similar to that used in our simulation. We compare our simulation results against that we find from real experimentation. The comparison reveals applicability of our cooling energy integration in SimGrid for simulating diversified distributed systems.

Secured aerial photography using Homomorphic Encryption

Aerial photography is fast becoming essential in scientific research that requires multi-agent system in several perspective and we proposed a secured system using one of the well-known public key cryptosystem namely NTRU that is somewhat homomorphic in nature. Here we processed images of aerial photography that were captured by multi-agents. The agents encrypt the images and upload those in the cloud server that is untrusted. Cloud computing is a buzzword in modern era and public cloud is being used by people everywhere for its shared, on-demand nature. Cloud Environment faces a lot of security and privacy issues that needs to be solved. This paper focuses on how to use cloud so effectively that there remains no possibility of data or computation breaches from the cloud server itself as it is prone to the attack of treachery in different ways. The cloud server computes on the encrypted data without knowing the contents of the images. After concatenation, encrypted result is delivered to the concerned authority where it is decrypted retaining its originality. We set up our experiment in Amazon EC2 cloud server where several instances were the agents and an instance acted as the server. We varied several parameters so that we could minimize encryption time. After experimentation we produced our desired result within feasible time sustaining the image quality. This work ensures data security in public cloud that was our main concern.

User interaction analysis to recommend suitable jobs in career-oriented social networking sites

Career-oriented social networking sites are very much useful for job seekers to find a suitable job and useful for recruiters as well to find the right candidate for a job. Job recommendation system helps job seekers to find appropriate jobs matching with their profile. So, it can be considered as recruiters approaching a suitable candidate whenever they have an appropriate job for them. In this paper, we present a research technique of developing a job recommendation system for the online job hunting websites to predict suitable job postings that are likely to be relevant to the user i.e., the job postings with which the users can possibly interact. Relevant jobs are those job postings on which a user may click, bookmark or reply to the recruiter. Here, we have considered all possible factors related to users as well as job items available in a publicly available partial big data set of a widely used international job hunting website. We have split the interaction data into training and test data for the purpose of evaluating our proposed system. After that we have developed an algorithm for job recommender system which can calculate similarity for user-user, item-item and hybrid of user-user and item-item perspective between training and test data set based on weighted scores using our proposed similarity computation algorithm for different jobs and users all the information present in the dataset. We have used Collaborative Filtering (CF) algorithm separately for user-user and item-item based approach and for hybrid approach, we have calculated the intersection between user-user and item-item based recommended list and select top-k job items as recommended lists from the intersection. After that, we have compared the predicted recommended list based on all three approaches with the actual list and made offline evaluation of the job recommender system accuracy based on obtained score. Finally we have found that hybrid approach performs better than user-user and item-item based approach for entire 90% sparsity of the training data.