Research

My research interest is broadly in the energy efficiency and power management of data centers. Data centers consume about 2% of the planet's electricity. In the U.S. alone, data centers consumed about 75 billion kWh electricity in 2020. My research focus is to explore ways to make data centers more energy efficient as well as to reduce their energy use. I have developed metrics to identify energy efficient servers for a data center.

I am also investigating ways to oversubscribe the data center power hierarchy as well as problems that arise in such data centers. Data centers are expensive infrastructures and cost of building a data center can range from $10 to $20 per watt (that puts the cost of building a 1 MW data center between $10 Million to $20 Million). Since servers do not consume their rated peak power all the time, data center owners are incentivized to oversubscribe (install more servers than allowed by the power capacity limit) their existing power infrastructure to increase data center power infrastructure utilization and to lower the Total Cost of Ownership (TCO). I have looked at how the server energy proportionality can effect possible safe power oversubscription of a data center. I have also investigated coordinated charging of distributed batteries in oversubscribed data centers.

I have done work in safe oversubscription of power infrastructure in multi-tenant data centers (MTDC). An MTDC is a data center where the operator owns the infrastructure (building, power, and cooling) and leases the facility to multiple tenants. Tenants, who pay a monthly lease bill, may have their own servers installed and in turn provide a service to their own customer. One important distinction is that, the operator does not have direct control over a tenant's servers or workload. This requires new solution and mechanisms, compared to operator owned data center, such that the operator can have coordination and distributed control of tenants in an MTDC.

Publications

Journal

1. S. Malla and K. Christensen, "The Effect of Server Energy Proportionality on Data Center Power Oversubscription," Future Generation Computer Systems Vol. 104, pp. 119-130, March 2020. (PDF | Link)
2. S. Malla and K. Christensen, "HPC in the Cloud: Performance Comparison of Function as a Service (FaaS) vs Infrastructure as a Service (IaaS)," Internet Technology Letters, Vol. 3, no. 1, 2020. (PDF | Link)
3. S. Malla and K. Christensen, "A Survey on Power Management Techniques for Oversubscription of Multi-Tenant Data Centers," ACM Computing Surveys 52, 1, Article 1, February 2019. (PDF | Link)

Conference

1. S. Malla, J. Wang, W. Hendrix, and K. Christensen, "Predicting Success for Computer Science Students in CS2 using Grades in Previous Courses," accepted to IEEE Frontiers in Education, October 2019. (PDF | Link)
2. S. Malla and K. Christensen, "Choosing the Best Server for a Data Center: The Importance of Workload Weighting," Proceedings of the IEEE International Performance Computing and Communications Conference, November 2018. (PDF | Link)
3. S. Malla, A. Tuladhar, G. J. Quadri, and P. Rosen, "Multi-Spectral Satellite Image Analysis for Feature Identification and Change Detection VAST Challenge 2017: Honorable Mention for Good Facilitation of Single Image Analysis," Proceedings of the IEEE Conference on Visual Analytics Science and Technology, pp. 205-206, October 2017. (PDF | Link)
4. G. J. Quadri, A. Tuladhar, S. Malla, and P. Rosen, "Visual Analytic Design for Characterizing Air-Sampling Sensor Performance and Operation," Proceedings of the IEEE Conference on Visual Analytics Science and Technology, pp. 217-218, October 2017. (PDF | Link)
5. A. Tuladhar, S. Malla, G. J. Quadri, and P. Rosen, "Data Aggregation and Visualization Technique for Traffic Sensor Data," Proceedings of the IEEE Conference on Visual Analytics Science and Technology, pp. 239-240, October 2017. (PDF | Link)
6. S. Malla and K. Christensen, "Reducing Power Use and Enabling Oversubscription in Multi-Tenant Data Centers Using Local Price," Proceedings of the IEEE International Conference on Autonomic Computing, pp. 161-167, July 2017. (PDF | Link)

Poster

1. S. Malla and K. Christensen, "Predicting Success for Computer Science Students," USF Graduate Student Research Symposium, March 2019. (PDF)
2. S. Malla and K. Christensen, "Using Local Power Price to Manage Multi-Tenant Data Center Performance and Energy," USF Graduate Student Research Symposium, November 2016. (PDF)

Research projects

Florida IT Pathways to Success

Under the guidance of Dr. Ken Christensen and Dr. Rafael Perez, I work as a graduate assistant for "Florida IT Pathways to Success" (Flit-Path) project. Flit-Path is an S-STEM project funded by National Science Foundation with the goal of providing scholarship, tutoring, and mentorship to academically talented but financially needy students at the Department of Computer Science and Engineering. This is a collaborative research effort between University of South Florida, Florida International University, and University of Central Florida.

VAST Challenge 2017

In summer of 2017, I along with my friends took part in the VAST Challenge under the guidance of Dr. Paul Rosen. VAST challenge is an annual competition where competing teams apply visual analytics on the provided data to solve the given problem. I worked on mini-challenge 3 which was about analyzing the given multi-spectral satellite image data to figure out the cause of decline in population of nesting pair of Pipit bird. We were required to complete this answer sheet along with an explanatory video. Our submission was awarded honorable mention for "Good Facilitation of Single Image Analysis" at the IEEE VIS 2017 conference held on October 1, 2017 in Phoenix, AZ.