Research Matters

Multi-cloud business applications without the lag or the overhead

Multi-cloud business applications without the lag or the overhead

The intercloud merging individual clouds into one seamless mass

Fourteen years ago, Amazon introduced its cloud-based solutions to the public, and the way software applications are delivered forever changed. With the widespread use of cloud computing, the number of clouds continued to grow, giving birth to what is now known as the intercloud.
The intercloud – or cloud of clouds – is the combination of many individual clouds into one seamless mass, where any single cloud is able to use resources beyond its reach. Some might think of an intercloud in relation to clouds, as the internet is in relation to Local Area Networks (LANs).
Multi-cloud computing, as is the next upcoming trend, will become a common practice in the next few years. This level of resource utilization opens up a whole new world of possibilities for distributed business applications – where software runs on multiple machines or servers.
A cloud-based application delivery process offers benefits that include simplified infrastructure, reduced costs, increased productivity, and improved end-user experience.
The Internet of Things (IoT) and Cyber-Physical Systems (CPS), for instance, are massively distributed by their very nature and could benefit greatly from multi-cloud delivery. Other businesses – such as banks, as well as financial, governmental, and educational institutions – can also benefit from a cloud-based deployment and will need to use a distributed multi-cloud computation infrastructure.
However, the problem with distributed cloud-based business applications is that some services need to gather, process, and forward data under near-real-time constraints, in addition to fault and performance issues potentially leading to interruptions in service (performance problem). On the operational expenditure front, Intercloud WAN links are expensive and have a high-cost bandwidth (cost problem).
Inspired by these challenges, Dr. Aiman Erbad and Dr. Mohammed Samaka from Qatar University, and Dr. Raj Jain from Washington University, St. Louis (WUSTL), developed new techniques for placing computations in a multi-cloud environment in an NPRP cycle 8 project (NPRP8-634-1-131) titled ‘Intercloud WAN Delivery Platform for distributed Business Applications.’
To solve these problems, the three professors designed and implemented several virtual-function and micro-services placement algorithms using integer linear programming, machine learning, and other techniques.
This gave birth to a platform that allows Internet Service Providers (ISPs) to find the optimal locations of new virtual network services based on their policies and communicate with the cloud management system of that cloud. Proper placement of the virtual application modules among geographically disparate clouds is important to ensure acceptable latency and throughput.
The project also proposed a fault detection and localization framework that allows service providers to predict faults in advance, and therefore avoid service interruptions. Furthermore, the team demonstrated the feasibility of supervised machine learning techniques for anomaly detection and categorization of attacks on business applications.
When combined, these technologies result in cost savings, higher performance, and added security for all distributed cloud-based applications.
According to Dr. Jain, the outcomes of the project will serve various industries and mega-projects that depend on a cloud computing infrastructure. “Qatar is actively working on several smart city projects and programs in order to meet the delivery of the 2022 FIFA World Cup. For reliability and scalability, the IT infrastructure, even though virtually centralized, should consist of a distributed multi-cloud environment”, he explained.
Dr. Jain also noted that ISPs and mobile telecommunication companies plan to use cloud computing to reduce their capital expenditure and operational expenditure by using virtualization, programmability, dynamic scaling, automation, and easy management of resources.
In Qatar, for example, the platform developed by the three professors can be utilized by the two main ISPs – Ooredoo and Vodafone – as well as by major Application Service Providers (ASPs) such as Hamad Medical Corporation, Qatar National Bank, Qatar Airways, Al Jazeera, and many more.
The promising advantages of the project have, to date, been published in eight papers, in addition to its results getting disseminated at a number of international conferences around the world.

Previous Article Media-use study shapes agendas in newsrooms, classrooms, and life-and-death affairs
Next Article Waste recycling: From research to practice
«March 2023»