I am in my third year of Computer Science Ph.D. study at Johns Hopkins University. I am very fortunate to be advised by Soudeh Ghorbani. My research interests span around designing operating systems and host networking componenets for datacenter applications. I'm currently investigating the sources and implications of data center traffic burstiness.
- Computer Science Ph.D., Johns Hopkins University
- Computer Science MEng., Johns Hopkins University
- Computer Software Engineering MSc., Iran University of Science and Technology
- Computer Software Engineering BSc., Iran University of Science and Technology
- Microbursts are one of the hidden culprits of data center network latency. We are investigating ways to identify, measure and mitigate them.
- Process scheduling is an old yet unsolved problem for data center hosts. With the emergence of kernel-bypass runtimes, new schemes are necessary to allow more efficient resource utilization while preserving the performance of latency-sensitive applications. We are currently investigting how micro-quanta scheduler tries to bring balance between fairness and responsiveness.
- Power management is another area of interest for data center operators. We recently proposed in-application process scheduling for latency-critical datacenter applications to leverage application knowledge for core management and ultimately smarter CPU power saving.
Erfan Sharafzadeh, Soudeh Ghorbani. 2021. Poster: A high-resolution study of data center traffic at its origin. In Proceedings of the 17th International Conference on emerging Networking EXperiments and Technologies (CoNEXT '21)
Sepehr Abdous*, Erfan Sharafzadeh*, and Soudeh Ghorbani. 2021. Burst-tolerant datacenter networks with Vertigo. In Proceedings of the 17th International Conference on emerging Networking EXperiments and Technologies (CoNEXT '21), *co-first authorsslides
Esmail Asyabi, Azer Bestavros, Erfan Sharafzadeh, Timothy Zhu. 2020. Peafowl: In-application CPU Scheduling to ReducePower Consumption of In-memory Key-Value Stores. In Proceedings of the ACM Symposium on Cloud Computing (SoCC '20)
Erfan Sharafzadeh, Seyed Alireza Sanaee Kohroudi, Esmail Asyabi, and Mohsen Sharifi. 2019. Yawn: A CPU Idle-state Governor for Datacenter Applications. In Proceedings of the 10th ACM SIGOPS Asia-Pacific Workshop on Systems (APSys '19). ACM, New York, NY, USA, 91-98. DOI: https://doi.org/10.1145/3343737.3343740
CTS: An Operating System CPU Scheduler to Mitigate Tail Latency for Latency-Sensitive Multi-Threaded Applications
Esmail Asyabi, Erfan Sharafzadeh, SeyedAlireza SanaeeKohroudi, Mohsen Sharifi, CTS: An operating system CPU scheduler to mitigate tail latency for latency-sensitive multi-threaded applications, Journal of Parallel and Distributed Computing, Available online 12 April 2018, ISSN 0743-7315, https://doi.org/10.1016/j.jpdc.2018.04.003.
- Shadow PC member, 2022 ACM European Conference on Computer Systems (EuroSys 2022).
- Shadow PC member, 2021 ACM European Conference on Computer Systems (EuroSys 2021).
- Design and development of "Shariffood", Sharif University of Technology dorms meal reservation portal.
- Infrastructure lead of IUST Cloud Computing Center: Iran University of Science and Technology IAAS Cloud platform.
- Fall 2016, Advanced Programming, IUST.
- Fall 2017, Operating Systems, IUST
- Spring 2018, Cloud Computing Environments, IUST
- Fall 2018, Operating Systems, IUST
- Spring 2020, Cloud Computing, JHU, course page
When I'm away from my PC, I play soccer (goalkeeper indoors, full back outdoors!), play video games (Fable and Far cry fan!), or leaf through my favorite books (I'm a Tolkien fan). Let's chat if our interests align ;)