LLM agents are increasingly used for networking tasks, but evaluating them remains hard because existing benchmarks are often static and manually curated. NetArena addresses this by generating executable network tasks dynamically through a state-action abstraction and emulator-backed evaluation on correctness, safety, and efficiency metrics. However, expanding NetArena to new benchmarks still requires significant expert effort. In this demonstration, we present a compiler layer and show how it adapts new source benchmarks to NetArena. The compiler extracts benchmark components such as state setup, fault injection, tool interfaces, and evaluation logic, then converts them into NetArena's intermediate representation for generating valid task variants. Our preliminary study shows that it can scale NIKA to at least 5× more instances than before1.

Federated Learning (FL) under distributed concept drift is a largely unexplored area. Although concept drift is itself a well-studied phenomenon, it poses particular challenges for FL, because drifts arise staggered in time and space (across clients). Our work is the first to explicitly study data heterogeneity in both dimensions. We first demonstrate that prior solutions to drift adaptation, with their single global model, are ill-suited to staggered drifts, necessitating multiple-model solutions. We identify the problem of drift adaptation as a time-varying clustering problem, and we propose two new clustering algorithms for reacting to drifts based on local drift detection and hierarchical clustering. Empirical evaluation shows that our solutions achieve significantly higher accuracy than existing baselines, and are comparable to an idealized algorithm with oracle knowledge of the ground-truth clustering of clients to concepts at each time step.

We leverage a novel telemetry source available in public clouds today: periodic summaries of every flow that enters or leaves any VM. A key aspect is that such telemetry can be collected transparently to customers and with minimal impact on their workloads. By consuming this telemetry, we show how one may realize complete and dynamic graphs of the communication inside cloud subscriptions. We describe novel analyses over these communication graphs with implications on network security and management.

Analyzing network topologies and communication graphs is essential in modern network management. However, the lack of a cohesive approach results in a steep learning curve, increased errors, and inefficiencies. In this paper, we present a novel approach that enables natural-language-based network management experiences, leveraging large language models (LLMs) to generate task-specific code from natural language queries. This method addresses the challenges of explainability, scalability, and privacy by allowing network operators to inspect the generated code, removing the need to share network data with LLMs, and focusing on application-specific requests combined with program synthesis techniques. We develop and evaluate a prototype system using benchmark applications, demonstrating high accuracy, cost-effectiveness, and potential for further improvements using complementary program synthesis techniques.

The barrier to entry for network operators to use machine learning (ML) is high for operators who are not ML experts. Automated machine learning (AutoML) promises operators the ability to train ML models without requiring the expertise of data scientists or the need to learn ML. However, AutoML today: (a) is black-box; and (b) does not allow operators to leverage domain expertise. We start this paper by describing our broader vision for a domain-customized AutoML platform for networking and propose a set of potential solutions to realize that vision. As the first step, we introduce our feedback solution for AutoML that allows domain experts (who are not experts in ML) to better understand how to improve the input data to AutoML in order to achieve better accuracy.

Continuously monitoring a wide variety of performance and fault metrics has become a crucial part of operating large-scale datacenter networks. In this work, we ask whether we can reduce the costs to monitor - in terms of collection, storage and analysis - by judiciously controlling how much and which measurements we collect. By positing that we can treat almost all measured signals as sampled time-series, we show that we can use signal processing techniques such as the Nyquist-Shannon theorem to avoid wasteful data collection. We show that large savings appear possible by analyzing tens of popular measurement systems from a production datacenter network. We also discuss some challenges that must be solved when applying these techniques in practice.

Many large-scale machine learning (ML) applications need to perform decentralized learning over datasets generated at different devices and locations. Such datasets pose a significant challenge to decentralized learning because their different contexts result in significant data distribution skew across devices/locations. In this paper, we take a step toward better understanding this challenge by presenting a detailed experimental study of decentralized DNN training on a common type of data skew: skewed distribution of data labels across devices/locations. Our study shows that: (i) skewed data labels are a fundamental and pervasive problem for decentralized learning, causing significant accuracy loss across many ML applications, DNN models, training datasets, and decentralized learning algorithms; (ii) the problem is particularly challenging for DNN models with batch normalization; and (iii) the degree of data skew is a key determinant of the difficulty of the problem. Based on these findings, we present SkewScout, a system-level approach that adapts the communication frequency of decentralized learning algorithms to the (skew-induced) accuracy loss between data partitions. We also show that group normalization can recover much of the accuracy loss of batch normalization.

Specialized on-chip accelerators are widely used to improve the energy efficiency of computing systems. Recent advances in memory technology have enabled near-data accelerators (NDAs), which reside off-chip close to main memory and can yield further benefits than on-chip accelerators. However, enforcing coherence with the rest of the system, which is already a major challenge for accelerators, becomes more difficult for NDAs. This is because (1) the cost of communication between NDAs and CPUs is high, and (2) NDA applications generate a lot of off-chip data movement. As a result, as we show in this work, existing coherence mechanisms eliminate most of the benefits of NDAs. We extensively analyze these mechanisms, and observe that (1) the majority of off-chip coherence traffic is unnecessary, and (2) much of the off-chip traffic can be eliminated if a coherence mechanism has insight into the memory accesses performed by the NDA.

Exploiting data locality in GPUs is critical to making more efficient use of the existing caches and the NUMA-based memory hierarchy expected in future GPUs. While modern GPU programming models are designed to explicitly express parallelism, there is no clear explicit way to express data locality-i.e., reuse-based locality to make efficient use of the caches, or NUMA locality to efficiently utilize a NUMA system. On the one hand, this lack of expressiveness makes it a very challenging task for the programmer to write code to get the best performance out of the memory hierarchy. On the other hand, hardware-only architectural techniques are often suboptimal as they miss key higher-level program semantics that are essential to effectively exploit data locality. In this work, we propose the Locality Descriptor, a crossl-ayer abstraction to explicitly express and exploit data locality in GPUs. The Locality Descriptor (i) provides the software a flexible and portable interface to optimize for data locality, requiring no knowledge of the underlying memory techniques and resources, and (ii) enables the architecture to leverage key program semantics and effectively coordinate a range of techniques (e.g., CTA scheduling, cache management, memory placement) to exploit locality in a programmer-transparent manner. We demonstrate that the Locality Descriptor improves performance by 26.6% on average (up to 46.6%) when exploiting reuse-based locality in the cache hierarchy, and by 53.7% (up to 2.8X) when exploiting NUMA locality in a NUMA memory system.

This paper makes a case for a new cross-layer interface, Expressive Memory (XMem), to communicate higher-level program semantics from the application to the system software and hardware architecture. XMem provides (i) a flexible and extensible abstraction, called an Atom, enabling the application to express key program semantics in terms of how the program accesses data and the attributes of the data itself, and (ii) new cross-layer interfaces to make the expressed higher-level information available to the underlying OS and architecture. By providing key information that is otherwise unavailable, XMem exposes a new, rich view of the program data to the OS and the different architectural components that optimize memory system performance (e.g., caches, memory controllers). By bridging the semantic gap between the application and the underlying memory resources, XMem provides two key benefits. First, it enables architectural/system-level techniques to leverage key program semantics that are challenging to predict or infer. Second, it improves the efficacy and portability of software optimizations by alleviating the need to tune code for specific hardware resources (e.g., cache space). While XMem is designed to enhance and enable a wide range of memory optimizations, we demonstrate the benefits of XMem using two use cases: (i) improving the performance portability of software-based cache optimization by expressing the semantics of data locality in the optimization and (ii) improving the performance of OS-based page placement in DRAM by leveraging the semantics of data structures and their access properties.

Through the last decade, we have witnessed a surge of Internet of Things (IoT) devices, and with that a greater need to choreograph their actions across both time and space. Although these two problems, namely time synchronization and localization, share many aspects in common, they are traditionally treated separately or combined on centralized approaches that results in an inefficient use of resources, or in solutions that are not scalable in terms of the number of IoT devices. Therefore, we propose D-SLATS, a framework comprised of three different and independent algorithms to jointly solve time synchronization and localization problems in a distributed fashion. The first two algorithms are based mainly on the distributed Extended Kalman Filter (EKF) whereas the third one uses optimization techniques. No fusion center is required, and the devices only communicate with their neighbors. The proposed methods are evaluated on custom Ultra-Wideband communication Testbed and a quadrotor, representing a network of both static and mobile nodes. Our algorithms achieve up to three microseconds time synchronization accuracy and 30 cm localization error.

3D-stacked memory devices with processing logic can help alleviate the memory bandwidth bottleneck in GPUs. However, in order for such Near-Data Processing (NDP) memory stacks to be used for different GPU architectures, it is desirable to standardize the NDP architecture. Our proposal enables this standardization by allowing data to be spread across multiple memory stacks as is the norm in high-performance systems without an MMU on the NDP stack. The keys to this architecture are the ability to move data between memory stacks as required for computation, and a partitioned execution mechanism that offloads memory-intensive application segments onto the NDP stack and decouples address translation from DRAM accesses. By enhancing this system with a smart offload selection mechanism that is cognizant of the compute capability of the NDP and cache locality on the host processor, system performance and energy are improved by up to 66.8% and 37.6%, respectively.

Pointer chasing is a fundamental operation, used by many important data-intensive applications (e.g., databases, key-value stores, graph processing workloads) to traverse linked data structures. This operation is both memory bound and latency sensitive, as it (1) exhibits irregular access patterns that cause frequent cache and TLB misses, and (2) requires the data from every memory access to be sent back to the CPU to determine the next pointer to access. Our goal is to accelerate pointer chasing by performing it inside main memory, thereby avoiding inefficient and high-latency data transfers between main memory and the CPU. To this end, we propose the In-Memory PoInter Chasing Accelerator (IMPICA), which leverages the logic layer within 3D-stacked memory for linked data structure traversal. This paper identifies the key design challenges of designing a pointer chasing accelerator in memory, describes new mechanisms employed within IMPICA to solve these challenges, and evaluates the performance and energy benefits of our accelerator. IMPICA addresses the key challenges of (1) how to achieve high parallelism in the presence of serial accesses in pointer chasing, and (2) how to effectively perform virtual-to-physical address translation on the memory side without requiring expensive accesses to the CPU's memory management unit. We show that the solutions to these challenges, address-access decoupling and a region-based page table, respectively, are simple and low-cost. We believe these solutions are also applicable to many other in-memory accelerators, which are likely to also face the two challenges. Our evaluations on a quad-core system show that IMPICA improves the performance of pointer chasing operations in three commonly-used linked data structures (linked lists, hash tables, and B-trees) by 92%, 29%, and 18%, respectively. This leads to a significant performance improvement in applications that utilize linked data structures - on a real database application, DBx1000, IMPICA improves transaction throughput and response time by 16% and 13%, respectively. IMPICA also significantly reduces overall system energy consumption (by 41%, 23%, and 10% for the three commonly-used data structures, and by 6% for DBx1000).

Main memory bandwidth is a critical bottleneck for modern GPU systems due to limited off-chip pin bandwidth. 3D-stacked memory architectures provide a promising opportunity to significantly alleviate this bottleneck by directly connecting a logic layer to the DRAM layers with high bandwidth connections. Recent work has shown promising potential performance benefits from an architecture that connects multiple such 3D-stacked memories and offloads bandwidth-intensive computations to a GPU in each of the logic layers. An unsolved key challenge in such a system is how to enable computation offloading and data mapping to multiple 3D-stacked memories without burdening the programmer such that any application can transparently benefit from near-data processing capabilities in the logic layer. Our paper develops two new mechanisms to address this key challenge. First, a compiler-based technique that automatically identifies code to offload to a logic-layer GPU based on a simple cost-benefit analysis. Second, a software/hardware cooperative mechanism that predicts which memory pages will be accessed by offloaded code, and places those pages in the memory stack closest to the offloaded code, to minimize off-chip bandwidth consumption. We call the combination of these two programmer-transparent mechanisms TOM: Transparent Offloading and Mapping. Our extensive evaluations across a variety of modern memory-intensive GPU workloads show that, without requiring any program modification, TOM significantly improves performance (by 30% on average, and up to 76%) compared to a baseline GPU system that cannot offload computation to 3D-stacked memories.

This paper introduces a new resource virtualization framework, Zorua, that decouples the programmer-specified resource usage of a GPU application from the actual allocation in the on-chip hardware resources. Zorua enables this decoupling by virtualizing each resource transparently to the programmer. The virtualization provided by Zorua builds on two key concepts - dynamic allocation of the on-chip resources and their oversubscription using a swap space in memory. Zorua provides a holistic GPU resource virtualization strategy, designed to (i) adaptively control the extent of oversubscription, and (ii) coordinate the dynamic management of multiple on-chip resources (i.e., registers, scratchpad memory, and thread slots), to maximize the effectiveness of virtualization. Zorua employs a hardware-software code-sign, comprising the compiler, a runtime system and hardware-based virtualization support. The runtime system leverages information from the compiler regarding resource requirements of each program phase to (i) dynamically allocate/deallocate the different resources in the physically available on-chip resources or their swap space, and (ii) manage the tradeoffbetween higher thread-level parallelism due to virtualization versus the latency and capacity overheads of swap space usage. We demonstrate that by providing the illusion of more resources than physically available via controlled and coordinated virtualization, Zorua offers several important benefits: (i) Programming Ease. Zorua eases the burden on the programmer to provide code that is tuned to efficiently utilize the physically available on-chip resources. (ii) Portability. Zorua alleviates the necessity of re-tuning an application's resource usage when porting the application across GPU generations. (iii) Performance. By dynamically allocating resources and carefully oversubscribing them when necessary, Zorua improves or retains the performance of applications that are already highly tuned to best utilize the hardware resources. The holistic virtualization provided by Zorua can also enable other uses, including fine-grained resource sharing among multiple kernels and low-latency preemption of GPU programs.

Long DRAM latency is a critical performance bottleneck in current systems. DRAM access latency is defined by three fundamental operations that take place within the DRAM cell array: (i) activation of a memory row, which opens the row to perform accesses; (ii) precharge, which prepares the cell array for the next memory access; and (iii) restoration of the row, which restores the values of cells in the row that were destroyed due to activation. There is significant latency variation for each of these operations across the cells of a single DRAM chip due to irregularity in the manufacturing process. As a result, some cells are inherently faster to access, while others are inherently slower. Unfortunately, existing systems do not exploit this variation.