National Institute of Technology Andhra Pradesh inks an MoU with LinkedIn

TADEPALLIGUDEM : National Institute of Technology Andhra Pradesh has signed a Memorandum of Understanding (MoU) with LinkedIn on 3rd January 2022. This will facilitate collaborative research and development between the two organizations.

The MoU was signed by Prof. C. S. P. Rao, Director, National Institute of Technology Andhra Pradesh, and Mr. Naga Krishna R. Sathvalli, Director, Business Systems Program, LinkedIn. Mr. Chidambaran Kollengode, Director, Data and AI Platforms division of LinkedIn spearheads this research collaboration on behalf of LinkedIn.

Under this MoU a research grant will be awarded to the principal investigator Dr. Karthick Seshadri, Assistant Professor, NIT Andhra Pradesh and the Co-Investigator Dr. Nagesh Bhattu Sristy, Assistant Professor, NIT Andhra Pradesh, for solving open research problems in enabling intelligent and elastic cloud computing.

Speaking about this initiative, Prof. C. S. P. Rao, Director, National Institute of Technology Andhra Pradesh, said, “The association of the two organisations would enable joint research in the niche area of intelligent cloud computing and in fostering research to address industry relevant problems by the academia.”

Further, Prof. C. S. P. Rao, said, “I appreciate the researchers at NIT Andhra Pradesh for providing an extensive training to their research team on Statistical Modeling, Cloud Computing and Machine Learning, to impart the necessary technical-know-how for meeting the objectives of the first MoU.”

This MoU is the second of its kind between the two organizations, which has been made possible after the researchers at NIT Andhra Pradesh showcased significant novel research outcomes as a deliverable of the first MoU which was signed in June 2020.

Dr. Karthick Seshadri, Assistant Professor, NIT Andhra Pradesh, said, “As an outcome of the first MoU with LinkedIn, we have set-up a cloud computing test-bed using the servers sponsored through funding by LinkedIn to run the experiments at NIT Andhra Pradesh premises. To resolve the open research problems faced in orchestrating an elastic cloud, we have used workloads sampled from a real time cloud environment to develop first-of-its-kind cloud workload characterization models and to subsequently utilize these models for capacity planning. The descriptive and predictive accuracies of the characterization models designed are better than the state-of-the-art models.”

Mr. Chidambaran Kollengode, Director, Data and AI Platforms division of LinkedIn, said, “Scaling is a big challenge in cloud environment and truly elastic model is typically viewed as a financial optimization problem; however, we have attempted to explore the technical implications underlying this problem. To address the problem of scaling, there is a need for an efficient workload characterization model. I am happy to remark that this collaboration has been very promising and fruitful in our steps to realize a truly elastic cloud.”

The researchers of NIT Andhra Pradesh and Mr. Chidambaran Kollengode have also devised an ‘Unsupervised model for profiling the workloads arriving at the data center and a supervised ensemble model for prediction of resource demands on cloud.’

The findings of this research work were accepted for publication in ‘Advances in Data Computing, Communication and Security,’ a Journal by Springer. Further, Dr. Karthick Seshadri and team have developed another efficient characterization model based on double embedded stochastic processes and their findings is in the pipeline to be published in the Journal of Supercomputing. This characterization model helps the data center administrators for capacity planning and assures proper resource provisioning. This efficient prediction model also helps in reduction of carbon emissions from data centers.

The objectives of the second MoU include the following:

(i) To nurture and create a talent-pool with a niche skillset in Cloud Computing and the broad domains of Machine Learning and Statistical Modeling for data analytics and training them to be industry-ready

(ii) To produce a Ph.D. student out of this project with a strong skillset to solve contemporary problems in Distributed Systems Engineering and Computing

(iii) To publish research articles in impact journals of high repute with the findings from this collaborative work

(iv) To inculcate the culture of organizing faculty development programs on Cloud Computing by sharing the findings of this project with the peer research groups and faculty across India

Further, Dr. Karthick Seshadri, said, “I thank Prof. C. S. P. Rao for his continuous encouragement and support in enabling us to execute high-quality and impactful research. Our research team is planning to extend the proposed framework to characterize the long and short-range dependencies in sequential workloads. The team will also explore the possibility of designing and prototyping a meta-statistical model for capacity planning. Our team also looks forward to evaluate the effectiveness of aforementioned technical approaches in reducing the spin-up time for stateful workloads.”

On the collaboration, Dr. Nagesh Bhattu Sristy, Assistant Professor, NIT Andhra Pradesh, said, “I am happy to collaborate with the Data and AI Platforms division of LinkedIn as one of the investigators of this research project and wish to carry out a high impact research to solve the complex engineering problems posed by LinkedIn in the domain of Cloud Computing.”