Software

Mutant (NSDI 2025)

Learning how to control congestion remains a challenge despite years of progress. Existing congestion control protocols have demonstrated efficacy within specific network conditions, inevitably behaving suboptimally or poorly in others. Machine learning solutions to congestion control have been proposed, though relying on extensive training and specific network configurations. In this paper, we loosen such dependencies by proposing textit{Mutant}, an online reinforcement learning algorithm for congestion control that adapts to the behavior of the best-performing schemes, outperforming them in most network conditions. Design challenges included determining the best protocols to learn from, given a network scenario, and creating a system able to evolve to accommodate future protocols with minimal changes. Our evaluation on real-world and emulated scenarios shows that Mutant achieves lower delays and higher throughput than prior learning-based schemes while maintaining fairness by exhibiting negligible harm to competing flows, making it robust across diverse and dynamic network conditions.

Mutant Github repository code

Neighborhhood Method (NM) Prototype

The NM is an optimal virtual path embedding method used in both control plane path finder protocols, such as virtual link embedding or NFV chain instantiation, and data plane path management protocols such as traffic steering. The prototype implementation augment the Floodlight Java controller and can be forked from this bitbucket repository (Oct 2016).

VINEA

VINEA is a VIrtual Network Embedding Architecture, part of my Ph.D thesis. The VINEA architecture prototype enables users to convert high-level policies into low-level virtual network embedding rules.

The VINEA prototype can be forked from this github repository.

Read More >

ProtoRINA

ProtoRINA is a Java prototype implementation of a clean-slate Recursive InterNetwork Architecture that is based on the fundamental principle that networking is inter-process communication (IPC). It recurses the IPC service over different scopes, allowing for better scalability, security, and manageability.

The RINA prototype (ProtoRINA) can be forked from this github repository.

Read More >

Alloy Model for Max-Consensus Auction (MCA) Protocols

In this project we propose a formal, machine-readable, Max-Consensus Auction model, encoded in the Alloy lightweight modeling language. The model consists of a network of agents applying the MCA mechanisms, instantiated with potentially different policies, and a set of predicates to analyze its convergence properties. Our model can be used to verify, with a “push-button” analysis, the convergence of the MCA mechanism to a conflict-free allocation of a wide range of policy instantiations.

The Alloy model (implemented by Saber Mirzaei) can be forked from this github repository.

Read More >

BUtorrent

BUtorrent is a file-sharing client that modifies the scheduling of the seed in the BitTorrent protocol. Our seed scheduling algorithm is based on a proportional-fair sharing approach, whereby pieces of the file with higher short-term demand, but lower long-term service rate, are served by the seed at higher priority. This ensures that, while meeting instantaneous need, pieces (replicas) are equally distributed within the network, thus improving the file-exchange rate among peers.

The instrumented (Python) BUTorrent client can be forked from this github repository.

Read More >

PREDA

Our PREDA system supports Predicate Routing in DTN-over-MANET networks. Predicate routing allows Delay-Tolerant-Network (DTN) users connected by an underlying Mobile Ad-hoc NETwork (MANET), to declaratively express high-level policy constraints on the routing of content. PREDA maps high-level constraints of DTN nodes to low-level routing predicates within the MANET nodes.

A VM running the PREDA Testbed with the demo showed at Mobicom can be downloaded at here.

Read More >

Trading Application Platform (TAP) ©

Exegy Trading Application Platform (TAP) allows raw market data feeds to be consumed and normalized on the same server that hosts a trading application. TAP is an embedded ticker plant product derived from Exegy’s performance know-how, elegant normalized data model, and proven feed handler technology. Available as a software library, TAP allows applications to leverage commodity hardware in order to achieve maximum deployment agility. Applications continue to use the Exegy Client API (XCAPI) to gain access to a broad spectrum of market data feeds and a rich set of market data views and features, including instant access to full depth-of-book views.