Searcharxiv⌕ Search

arXiv subjects

Gaurav Mitra

Publications and source records attributed to Gaurav Mitra.

5 recordsLinked to original sources

FIRE: Failure Isolation and Reliability for Production-Connected Experiments at Scale

Pre-production changes are usually tested in isolated environments to prevent their side effects from reaching production. Some record-and-replay platforms allow a production-connected mode: candidate changes receive recorded production requests while selected downstream calls reach production dependencies. This improves fidelity before deployment, but it also introduces reliability risks. At hyperscale, experimental traffic can fan out across shared services, trigger latent defects, or send malformed input to systems serving user requests. In this paper, we show how production-connected experimentation can operate reliably in a large-scale recommendation serving system. To contain downstream effects without sacrificing realistic dependency behavior, the platform uses three complementary guardrails. Isolated tiers redirect replay traffic away from shared dependencies where replication is practical. Dependency monitors observe downstream health and experiment-attributed errors and terminate experiments after sustained violations. Traffic pinning limits experimental calls to a designated subset of downstream workers when full replication would require scarce capacity. Isolated ranking tiers process billions of replayed requests per day using under 1% of the corresponding serving capacity. Over 90 days, these tiers experienced more than 200,000 abnormal worker exits without directly affecting production-serving workers. During a separate three-month window, dependency monitors cancelled approximately 5% of experiments for the measured ranking service and accounted for 87% of its cancellations. Dependency-aware triggering also reduced code-review experiment executions by 25%. These results show that high-fidelity, production-connected experimentation can operate at scale when downstream effects are isolated where possible, bounded where necessary, and continuously monitored and attributed.

cs.SE↗

Diviner: Agentic RCA with Human-Governed Adaptation for Large-Scale Infrastructure

Root-cause analysis (RCA) in large-scale distributed systems is challenging because a single symptom can cross service, data, configuration, compute, and organizational boundaries. Responders must assemble incomplete evidence and coordinate experts who hold different parts of the causal path. Generic language-model agents do not solve this problem through tool access alone: they can use stale or mistimed context, interpret failed queries as absent symptoms, report unsupported system state, and overload observability services designed for human-paced use. We present Diviner (Domain-Informed Verifiable Incident Navigation through Evidence and Review), an agentic RCA system operating at Meta and organized around a four-stage reliability lifecycle. Encode combines a large corpus of reviewed incidents with service-owner expertise to produce scoped, versioned skills, workflows, typed tools, and an incident knowledge graph. Execute uses an adaptive Lead and context-isolated workers to apply relevant knowledge to point-in-time evidence and maintain evidence-linked hypotheses. Evaluate compares saved investigation trajectories with responder-reviewed outcomes to measure exact cause, useful direction, and Groundedness. Adapt uses recurring gaps and on-call feedback to derive focused operational- knowledge updates, admitting them only after replay, quality checks, and owner approval.

cs.AI↗

Simulation and Network Assembly Pipelines for Dynamically Bonded Soft Materials

Soft materials linked by reversible covalent or supramolecular bonds form a diverse class of assemblies with promising applications from nanoscience to medicine. Experiments typically probe bulk phase behavior and rheology, but it remains difficult to measure how microscopic bonding kinetics and the mechanics of the constituent elements give rise to bulk properties. Coarse-grained molecular dynamics (MD) simulations can bridge these scales, but most simulation approaches do not control individual bond kinetics, and those that do were mostly developed for bespoke applications that do not readily generalize. Here we present pySNAP (Simulation and Network Assembly Pipelines), a modular open source Python platform that integrates tunable dynamic bonding with a workflow, template, and analysis setup, so that users can study a wide range of systems with only small changes to input files. The platform is built on the GPU-accelerated HOOMD-blue MD engine and integrates DyBond, a GPU-accelerated plugin that forms and breaks bonds consistent with an equilibrium distribution and supports bonding between multiple types of partner species. Around this core, the snap_simulate package compiles a directory of parameter files into a HOOMD-blue simulation, and the snap_workflow package orchestrates the resulting parameter sweeps across workstations and high-performance computing schedulers. We describe the theory behind simulated dynamic bonding and how to use the package, from setting up a parameter sweep to analyzing its results, and demonstrate the framework on a diverse range of dynamically bonded systems, showing that it accommodates distinct interaction mechanisms, geometries, and physical scenarios within a unified workflow, while enabling both reproduction of existing models and rapid construction of more complex composite systems.

cond-mat.soft↗

A Coarse-Grained Simulation Model for Self-Assembly of Liquid Droplets Featuring Explicit Mobile Binders

Colloidal particles with mobile binding molecules constitute a powerful platform for probing the physics of self-assembly. Binding molecules are free to diffuse and rearrange on the surface, giving rise to spontaneous control over the number of droplet-droplet bonds, i.e., valence, as a function of the concentration of binders. This type of valence control has been realized experimentally by tuning the interaction strength between DNA-coated emulsion droplets. Optimizing for valence two yields droplet polymer chains, termed `colloidomers', which have recently been used to probe the physics of folding. To understand the underlying self-assembly mechanisms, here we present a coarse-grained molecular dynamics (CGMD) model to study the self-assembly of this class of systems using explicit representations of mobile binding sites. We explore how valence of assembled structures can be tuned through kinetic control in the strong binding limit. More specifically, we optimize experimental control parameters to obtain the highest yield of long linear colloidomer chains. Subsequently tuning the dynamics of binding and unbinding via a temperature-dependent model allows us to observe the heptamer chain collapse into all possible rigid structures, in good agreement with recent folding experiments. Our CGMD platform and dynamic bonding model (implemented as an open-source custom plugin to HOOMD-Blue) reveal the molecular features governing the binding patch size and valence control, and opens the study of pathways in colloidomer folding. This model can therefore guide programmable design in experiments.

cond-mat.soft↗

Programming the Adapteva Epiphany 64-core Network-on-chip Coprocessor

In the construction of exascale computing systems energy efficiency and power consumption are two of the major challenges. Low-power high performance embedded systems are of increasing interest as building blocks for large scale high- performance systems. However, extracting maximum performance out of such systems presents many challenges. Various aspects from the hardware architecture to the programming models used need to be explored. The Epiphany architecture integrates low-power RISC cores on a 2D mesh network and promises up to 70 GFLOPS/Watt of processing efficiency. However, with just 32 KB of memory per eCore for storing both data and code, and only low level inter-core communication support, programming the Epiphany system presents several challenges. In this paper we evaluate the performance of the Epiphany system for a variety of basic compute and communication operations. Guided by this data we explore strategies for implementing scientific applications on memory constrained low-powered devices such as the Epiphany. With future systems expected to house thousands of cores in a single chip, the merits of such architectures as a path to exascale is compared to other competing systems.

cs.AR↗