Ranysha Ware

I am an Assistant Professor in the Computer Science Department at Swarthmore College where I teach systems courses. In my research, I build tools to make evaluating and understanding the Internet and networked systems easier and more accurate.

Previously I was a Postdoctoral Teaching Fellow in CMU’s Computer Science Department. I completed my PhD at CMU also in CSD where I was co-advised by Professor Justine Sherry and Professor Srinivasan Seshan. My thesis was about developing a new methodology for determining if a new congestion control algorithm is safe to deploy in the Internet today.

I earned my M.S. in Computer Science from University of Massachusetts Amherst where I worked with Professor Charles Weems in the Architecture and Language Implementation Group. I earned my B.S. in Computer Science from The State University of New York at New Paltz. Prior to CMU, I was Associate Technical Staff at MIT Lincoln Laboratory in the Cyber Analytics and Decision Systems Group.

My dissertation won the SIGCOMM Doctoral Dissertation Award. In addition, I was a Facebook Emerging Scholar, a two-time receipient of the National GEM Consortium Fellowship and was named one of SUNY New Paltz’s 40 under Forty Alumni in 2017.

Here is my CV.

teaching

CS 43: Computer Networks

Semester: Fall 2025

This course will introduce fundamental ideas in computer networks. The structure of this course may be different from many other CS courses at Swarthmore. We’ll be using a teaching model called peer instruction, which places a strong emphasis on classroom discussion and student interaction.

news

May 2025

I won the SIGCOMM Doctoral Dissertation Award for my PhD thesis Battle for Bandwidth: On The Deployability of New Congestion Control Algorithms!

Feb 2025

I accepted a postion as an Assistant Professor of Computer Science at Swarthmore College starting Fall 2025.

Jan 2025

I am teaching 15-213 at CMU for Spring 2024: 15-213: Introduction to Computer Systems.

Aug 2024

I am teaching 15-110 at CMU for Fall 2024: 15-110: Principles of Computing.

Jul 2024

Our paper Reverse-Engineering Congestion Control Algorithm Behavior was accepted to IMC 2024.

publications

Congestion Safety Audits for Internet Services Sherry, Justine, Ware, Ranysha, Suizo, Isabel, Seshan, Srinivasan, and Philip, Adithya Abraham TPRC-53 [Abstract] [PDF]

When two family members sharing a home Internet connection simultaneously log in to their online applications of choice, how well do their services perform? Will one videogaming child experience timeouts and stalls because their sibling is also streaming a popular TV show? Will a teleconferencing mother have her calls drop because her spouse chooses just this moment to install a software update? When multiple Internet users share a bottleneck network link and overload it, the link is said to become congested. Bandwidth – the volume of data that can be transmitted per second – must be shared between users. If the link is overloaded, then latency – the amount of time it takes one bit of information to travel from sender to receiver – will increase, and packets of data can be dropped entirely by the network. Most Internet service providers do little to police the impact of congestion between users, simply forwarding packets in the order that they arrive, “first come, first serve.” This policy leaves the responsibility of mitigating the effects of congestion to the senders themselves. Internet applications use algorithms called Congestion Control Algorithms (CCAs) to carefully speed up and slow down their transmission to avoid congestion. CCAs aim to share available bandwidth equally between connections, keep latencies low, and avoid packet loss. Unfortunately, recent research shows that the success of applications in equitably mitigating congestion is mixed at best. In a recent study of widely deployed services, it was found that one file-distribution application consumed well over its fair share of bandwidth, causing other concurrently running applications to obtain as little as 16% of their fair share of bandwidth. In the same study, it was found that a popular video streaming service could lead to stalls for competing teleconferencing applications (even though other video streaming services did not cause this problem). In this paper, we explore how to evaluate whether Internet applications are friendly or fair to other applications sharing the same bottleneck link. Taking inspiration from the security community, we propose congestion safety audits, a methodology for analyzing an Internet service to identify common-case pitfalls in congested environments. Much like it is impossible to prove that an Internet application is “definitely secure”, there exists no approach to guarantee that an application will always share bandwidth with other applications or never lead to packet loss or stalls for competing traffic. Security audits allow inspectors to perform a suite of tests and evaluations to ensure that due diligence has been carried out to attempt to secure an application. We propose that service operators can similarly use a standardized suite of tests to verify that their applications meet common-sense guidelines for performance under competition. With a standardized process for congestion safety auditing, companies can evaluate their own 1 applications to avoid bad publicity. Or, in a world in which regulators take an interest in behavior under competition, congestion safety audits might provide validation that operators have followed due diligence in avoiding overly aggressive behavior on the Internet

PhD Thesis: Battle for Bandwidth: On The Deployability of New Congestion Control Algorithms [PDF]

Reverse-Engineering Congestion Control Algorithm Behavior Ferreira, Margarida, Ware, Ranysha, Kothari, Yash, Lynce, Inês, Martins, Ruben, Narayan, Akshay, and Sherry, Justine In Proceedings of the 2024 ACM on Internet Measurement Conference (IMC ’24) [Abstract] [PDF]

The rise of proprietary and novel congestion control algorithms (CCAs) opens questions about the future of Internet utilization, latency, and fairness. However, fully analyzing how novel CCAs impact these properties requires understanding the inner workings of these algorithms. We thus aim to reverse-engineer deployed CCAs’ behavior from collected packet traces to facilitate analyzing them. We present Abagnale, a program synthesis pipeline that helps users automate the reverse-engineering task. Using Abagnale, we discover simple expressions capturing the behavior of 9 of the 16 CCAs distributed with the Linux kernel and analyze 7 CCAs from a graduate networking course.

CCAnalyzer: An Efficient and Nearly-Passive Congestion Control Classifier Ware, Ranysha, Philip, Adithya Abraham, Hungria, Nicholas, Kothari, Yash, Sherry, Justine, and Seshan, Srinivasan In Proceedings of the ACM SIGCOMM 2024 Conference (SIGCOMM ’24) [Abstract] [PDF]

We present CCAnalyzer, a novel classifier for deployed Internet congestion control algorithms (CCAs) which is more accurate, more generalizable, and more human-interpretable than prior classifiers. CCAnalyzer requires no knowledge of the underlying CCA algorithms, and it can identify when a CCA is novel - i.e. not in the training set. Furthermore, CCAnalyzer can cluster together servers it believes use the same novel/unknown algorithm. CCAnalyzer correctly identifies all 15 of the default Internet CCAs deployed with Linux, including BBRv1, which no existing classifier can do. Finally, CCAnalyzer can classify server CCAs while being as efficient or better than prior approaches in terms of bytes transferred and runtime. We conduct a measurement study using CCAnalyzer measuring the CCA for 5000+ websites. We find widespread deployment of BBRv1 at large CDNs, and demonstrate how our clustering technique can detect deployments of new algorithms as it discovers BBRv3 although BBRv3 is not in its training set.

Prudentia: Findings of an Internet Fairness Watchdog Philip, Adithya Abraham, Athapathu, Rukshani, Ware, Ranysha, Mkocheko, Fabian Francis, Schlomer, Alexis, Shou, Mengrou, Meng, Zili, Seshan, Srinivasan, and Sherry, Justine In Proceedings of the ACM SIGCOMM 2024 Conference (SIGCOMM ’24) [Abstract] [PDF]

With the rise of heterogeneous congestion control algorithms and increasingly complex application control loops (e.g. adaptive bitrate algorithms), the Internet community has expressed growing concern that network bandwidth allocations are unfairly skewed, and that some Internet services are ’winners’ at the expense of ’losing’ services when competing over shared bottlenecks. In this paper, we provide the first study of fairness between live, end-to-end services with distinct workloads. Rather than focusing on individual components of an application stack (e.g., studying the fairness of an individual congestion control algorithm), we want to provide a direct study over real-world deployed applications. Among our findings, we observe that services typically achieve less-than-fair outcomes: on average, the ’losing’ service achieves only 72% of its max-min fair share of link bandwidth. We also find that some services are significantly more contentious than others: for example, one popular file distribution service causes competing applications to obtain as low as 16% of their max-min fair share of bandwidth when competing in a moderately-constrained setting.

Revisiting TCP Congestion Control Throughput Models & Fairness Properties at Scale Philip, Adithya Abraham, Ware, Ranysha, Athapathu, Rukshani, Sherry, Justine, and Sekar, Vyas In Proceedings of the 21st ACM Internet Measurement Conference (IMC ’21) [Abstract] [PDF]

Much of our understanding of congestion control algorithm (CCA) throughput and fairness is derived from models and measurements that (implicitly) assume congestion occurs in the last mile. That is, these studies evaluated CCAs in "small scale" edge settings at the scale of tens of flows and up to a few hundred Mbps bandwidths. However, recent measurements show that congestion can also occur at the core of the Internet on inter-provider links, where thousands of flows share high bandwidth links. Hence, a natural question is: Does our understanding of CCA throughput and fairness continue to hold at the scale found in the core of the Internet, with 1000s of flows and Gbps bandwidths? Our preliminary experimental study finds that some expectations derived in the edge setting do not hold at scale. For example, using loss rate as a parameter to the Mathis model to estimate TCP NewReno throughput works well in edge settings, but does not provide accurate throughput estimates when thousands of flows compete at high bandwidths. In addition, BBR - which achieves good fairness at the edge when competing solely with other BBR flows - can become very unfair to other BBR flows at the scale of the core of the Internet. In this paper, we discuss these results and others, as well as key implications for future CCA analysis and evaluation.

Beyond Jain’s Fairness Index: Setting the Bar For The Deployment of Congestion Control Algorithms Ware, Ranysha, Mukerjee, Matthew K., Seshan, Srinivasan, and Sherry, Justine In Proceedings of the 18th ACM Workshop on Hot Topics in Networks (HotNets ’19) [Abstract] [PDF] [Slides] [Video]

The Internet community faces an explosion in new congestion control algorithms such as Copa, Sprout, PCC, and BBR. In this paper, we discuss considerations for deploying new algorithms on the Internet. While past efforts have focused on achieving ‘fairness’ or ‘friendliness’ between new algorithms and deployed algorithms, we instead advocate for an approach centered on quantifying and limiting harm caused by the new algorithm on the status quo. We argue that a harm-based approach is more practical, more future proof, and handles a wider range of quality metrics than traditional notions of fairness and friendliness.

Modeling BBR’s Interactions with Loss-Based Congestion Control Ware, Ranysha, Mukerjee, Matthew K., Seshan, Srinivasan, and Sherry, Justine In Proceedings of the Internet Measurement Conference (IMC ’19) [Abstract] [PDF] [Slides] [Video]

BBR is a new congestion control algorithm (CCA) deployed for Chromium QUIC and the Linux kernel. As the default CCA for YouTube (which commands 11+% of Internet traffic), BBR has rapidly become a major player in Internet congestion control. BBR’s fairness or friendliness to other connections has recently come under scrutiny as measurements from multiple research groups have shown undesirable outcomes when BBR competes with traditional CCAs. One such outcome is a fixed, 40% proportion of link capacity consumed by a single BBR flow when competing with as many as 16 loss-based algorithms like Cubic or Reno. In this short paper, we provide the first model capturing BBR’s behavior in competition with loss-based CCAs. Our model is coupled with practical experiments to validate its implications. The key lesson is this: under competition, BBR becomes window-limited by its ‘inflight cap’ which then determines BBR’s bandwidth consumption. By modeling the value of BBR’s in-flight cap under varying network conditions, we can predict BBR’s throughput when competing against Cubic flows with a median error of 5%, and against Reno with amedian of 8%.

Dr. Ranysha Ware

Swarthmore College. Teaching the youth.

teaching

CS 43: Computer Networks

news

publications