Overview

Cloud computing is the mainstream delivery of on-demand and easy-to-use services such as computing, networking, storage, databases, and software over the Internet. It provides a pay-per-usage model in which consumers typically pay for cloud services they use. Cloud computing comes into three models: public, private and hydrid. A public cloud provides computing, storage and networking resources to the general public over Internet while a private cloud facilitates resources of on-premises infrastructure for the dedicated use of a specific organization. A hybrid cloud is a seamless integration of public and private clouds to take the best of both worlds. It enables cloud bursting in which applications initially leverage private cloud and burst into a public cloud when private resources are not enough to provision under spiking workload. A hybrid cloud significantly benefits its owner in terms of security in compliance with the location of sensitive data, availability, reliability, and cost reduction. However, construction of hybrid clouds and deployment of applications on top of them are not trivial tasks. [what you are currenlty working on …]

Publications

• TBA

Project Members

• TBA

Overview

Recently, geo-distributed data analytics is becoming a well-known method to process and manage the big data generated from globally distributed data. The ability to analyze and process geo-distributed data has become an important and challenging mission in many domains. Many applications need to process and analyze a massive amount of geo-distributed data. For instance, a bioinformatics application that analyzes existing genomes in different laboratories, a system to monitor, analyze log files from distributed servers, or any recommendation system. Today, the well-known cloud providers unify the management of infrastructure and application across public and on-premises. AWS, Google, and Azure Stack hyper-converged infrastructure are examples of hybrid cloud solutions that seamlessly allow users to exploit native public cloud services on premises in a variety of form factors with the help of public cloud APIs. The widespread usage of resource-constrained mobile devices and mobile edge objects including resource-moderated incentive academia and industries sector to investigate the management of fragmented resources between connected between mobile devices/objects in cloud environments. Our primary focus of fragmented hybrid clouds is to enable distributed databases and big data analytical frameworks to connect with a single computing service, which underneath leverage several Geo-distributed private cloud services.

Publications

• TBA

Project Members

• TBA

Overview

The process of deploying, configuring and evaluating a cloud infrastructure is a high technical task required a user with a good understanding of networking and cloud computing. In addition, the software installation and configuration processes in cluster-based application like Hadoop, Spark, Cassandra, etc. always requires a huge amount of efforts.

CREST Cloud Tool provides a less-technical, user-friendly and intuitive web-based application for users with none-or-less cloud computing knowledge:

1. To automate the process of deploying, configuring cloud infrastructures.

2. To automate the process of installing, configuring, evaluating cluster-based applications on cloud infrastructures.

3. To be used for experimentation related to cloud and big data.

4. To collaborate on sharable infrastructure and applications.

5. To manage and monitor the cloud infrastructures of an organization.

Project Members

• Faheem Ullah
• Yaser Mansouri
• Victor Prokhorenko
• Trung Ky Moc

Overview

With the rapid increase in digital technology, the scale of data available to human society is proliferating and we have moved from collecting small samples of clean data to collecting large volume of structured and unstructured streaming data. This gave rise to Big Data, which refers to large or complex datasets which are beyond the limits of traditional databases to capture, manage or process with low latency. However, with the availability of powerful processing hardware, huge inexpensive memories, and cloud computing has enabled big data analytics to examine large amounts of data to find insights, derive conclusions, and to support data-driven decision making. By taking advantage of these technologies, we have developed the CREST big data platform which allows users to analyse data effortlessly without having the hassle of assembling or configuring of sophisticated hardware or software.

Our platform provides a comprehensive, flexible, and scalable ecosystem of frameworks and functionalities that allows capturing, processing, analysis, and visualization of large volumes of data. By providing data analytics as service, it allows analysts, researchers, and business users to make better and faster decisions using data that was previously inaccessible or unusable. By utilizing advanced analytics techniques such as machine learning, deep learning, predictive analytics, classification and pattern mining, the users can gain valuable insights from previously unusable data sources independently or together with other previously existing data.

Project Members

• Faheem Ullah
• Anjitha Sreekumar

Overview

Big data technologies have gained tremendous popularity over the last few years, with applications in several domains such as healthcare, business analytics and smart environments. In the cyber security domain, the traditional cyber security solutions such as Intrusion Detection System (IDS) are unable to cope up with the high Volume, Velocity, and Variety, known as the 3Vs of security event data. By incorporating big data technologies with cyber security gave rise to Big Data Cyber Security Analytics (BDCA) system which has provided means for collecting, storing, and analysing large volume of security events data to monitor and protect organizational networks, computers, and data from unauthorized access, damage, or attack. Working in this area for the last 4-5 years, we have developed and evaluated three adaptation approaches: ADABTics, QuickAdapt, and SCALER.

• ADABTics combines component-based architecture with architecture-driven adaptation to ensure optimal performance.

• QuickAdapt exploits the component based BDCA architecture to add or drop components at runtime to allow quick adaptation to the changes in the security events data. Instead of focusing on architecture.

• SCALER focusses on parameters to tune them as per the operating environment in order to ensure that the BDCA system scales optimally.

Publications

ICSA 2019, ICECCS 2019, and JNCA 2021

Project Members

• Faheem Ullah