Sub-millisecond Latency Research Articles

Emulation techniques for scenario and classical control design of tokamak plasmas

The optimization of scenarios and design of real-time-control in tokamaks, especially for machines still in design phase, requires a comprehensive exploration of solutions to the Grad–Shafranov (GS) equation over a high-dimensional space of plasma and coil parameters. Emulators can bypass the numerical issues in the GS equation, if a large enough library of equilibria is available. We train an ensemble of neural networks to emulate the typical shape-control targets (separatrix at midplane, X-points, divertor strike point, flux expansion, and poloidal beta) as a function of plasma parameters and active coil currents for the range of plasma configurations relevant to spherical tokamaks with a super-X divertor, with percent-level accuracy. This allows a quick calculation of the classical-control shape matrices, potentially allowing real-time calculation at any point in a shot with submillisecond latency. We devise a hyperparameter sampler to select the optimal network architectures and quantify uncertainties on the model predictions. To generate the relevant training set, we devise a Markov-chain Monte Carlo algorithm to produce large libraries of forward Grad–Shafranov solutions without the need for user intervention. The algorithm promotes equilibria with desirable properties, while avoiding parameter combinations resulting in problematic profiles or numerical issues in the integration of the GS equation.

Hybrid fault tolerance in distributed in-memory storage systems

An in-memory storage system provides submillisecond latency and improves the concurrency of user applications by caching data into memory from external storage. Fault tolerance of in-memory storage systems is essential, as the loss of cached data requires access to data from external storage, which evidently increases the response latency. Typically, replication and erasure code (EC) are two fault-tolerant schemes that pose different trade-offs between access performance and storage usage. To help make the best performance and space trade-off, we design ElasticMem, a hybrid fault-tolerant distributed in-memory storage system that supports elastic redundancy transition to dynamically change the fault-tolerant scheme. ElasticMem exploits a novel EC-oriented replication (EOR) that carefully designs the data placement of replication according to the future data layout of EC to enhance the I/O efficiency of redundancy transition. ElasticMem solves the consistency problem caused by concurrent data accesses via a lightweight table-based scheme combined with data bypassing. It detects corelated read and write requests and serves subsequent read requests with local data. We implement a prototype that realizes ElasticMem based on Memcached. Experiments show that ElasticMem remarkably reduces the time of redundancy transition, the overall latency of corelated concurrent data accesses, and the latency of single data access among them.

Message Authentication and Provenance Verification for Industrial Control Systems

Successful attacks against industrial control systems (ICSs) often exploit insufficient checking mechanisms. While firewalls, intrusion detection systems, and similar appliances introduce essential checks, their efficacy depends on the attackers’ ability to bypass such middleboxes. We propose a provenance solution to enable the verification of an end-to-end message delivery path and the actions performed on a message. Fast and flexible provenance verification (F2-Pro) provides cryptographically verifiable evidence that a message has originated from a legitimate source and gone through the necessary checks before reaching its destination. F2-Prorelies on lightweight cryptographic primitives and flexibly supports various communication settings and protocols encountered in ICS thanks to its transparent, bump-in-the-wire design. We provide formal definitions and cryptographically prove F2-Pro’s security. For human interaction with ICS via a field service device, F2-Profeatures a multi-factor authentication mechanism that starts the provenance chain from a human user issuing commands. We compatibility tested F2-Proon a smart power grid testbed and reported a sub-millisecond latency overhead per communication hop using a modest ARM Cortex-A15 processor.

Offline and Online Algorithms for SSD Management

Flash-based solid-state drives (SSDs) are a key component in most computer systems, thanks to their ability to support parallel I/O at sub-millisecond latency and consistently high throughput. At the same time, due to the limitations of the flash media, they perform writes out-of-place, often incurring a high internal overhead which is referred to as write amplification. Minimizing this overhead has been the focus of numerous studies by the systems research community for more than two decades. The abundance of system-level optimizations for reducing SSD write amplification, which is typically based on experimental evaluation, stands in stark contrast to the lack of theoretical algorithmic results in this problem domain. To bridge this gap, we explore the problem of reducing write amplification from an algorithmic perspective, considering it in both offline and online settings. In the offline setting, we present a near-optimal algorithm. In the online setting, we first consider algorithms that have no prior knowledge about the input and show that in this case, the greedy algorithm is optimal. Then, we design an online algorithm that uses predictions about the input. We show that when predictions are relatively accurate, our algorithm significantly improves over the greedy algorithm. We complement our theoretical findings with an empirical evaluation of our algorithms, comparing them with the state-of-the-art scheme. The results confirm that our algorithms exhibit an improved performance for a wide range of input traces.

Snout: A Middleware Platform for Software-Defined Radios

The plethora of Internet of Things (IoT) protocols and the upcoming availability of new spectrum bands for wirelessly connected devices have made software-defined radio (SDR) technology increasingly useful to interact with radio-based communication. While SDR-based tools have grown in popularity in recent years due to their flexibility and adaptability towards new protocols, SDR software interfaces remain highly complex and technical, and inherently require specialized skillsets in digital signal processing (DSP) to operate. To address this problem, we present Snout, an SDR middleware platform that encapsulates and abstracts much of the current complexity in SDR toolchains. This allows SDR developers to create wireless networking applications usable by a wide range of users. For instance, network security professionals can monitor the IoT landscape across multiple protocols without needing to interact with the underlying software-defined DSP. Snout implements interfaces with common network analysis tools to allow for integration with traditional network security solutions, facilitating use cases such as traffic analysis or rogue device detection. Its software architecture enables scalability in terms of protocols and processing by modularizing the signal processing pipeline. To demonstrate Snout’s capabilities, we show how it can encapsulate GNU Radio flowgraphs, facilitate simultaneous multi-protocol scanning, and convert existing SDR-based protocol implementations into fully contained applications. We further demonstrate how Snout can handle GNU Radio flowgraphs with other signal processing software simultaneously. Through extensive experiments, we demonstrate that Snout incurs limited CPU performance overhead below 4% and a memory footprint below 100MB, and handles large amounts of events with sub-millisecond latency.

XURLLC in 6G with meshed RAN

5G Ultra-Reliable Low Latency Communications Technology (URLLC) will not be able to provide extremely reliable low latency services to the complex networks in 6G. Moreover, URLLC that began with 5G has to be refined and improved in 6G to provide xURLCC (extreme URLCC) with sub-millisecond latency, for supporting diverse mission-critical applications. This paper aims to highlight the importance of peer-to-peer mesh connectivity for services that require xURLLC. Deploying mesh connectivity among RAN nodes would add significant value to the current 5G New Radio (5G NR) enabling 6G to increase flexibility and reliability of the networks while reducing the inherent latency introduced by the core network. To provide a mesh connectivity in RAN, the nodes should be able to communicate with each other directly and be independent from the mobile core network so that data can be directly exchanged between base stations (gNBs) whereas certain aspects of signalling procedure including data session establishment will be managed by RAN itself. In this paper, we introduce several architectural choices for a mesh network topology that could potentially be crucial to a number of applications. In addition, three possible options to create mesh connectivity in RAN are provided, and their pros and cons are discussed in detail.

IBalancer: Load-Aware in-Server Flow Scheduling for Sub-Millisecond Tail Latency

iBalancer: Load-Aware in-Server Flow Scheduling for Sub-Millisecond Tail Latency

An Enterprise-Grade Open-Source Data Reduction Architecture for All-Flash Storage Systems

All-flash storage (AFS) systems have become an essential infrastructure component to support enterprise applications, where sub-millisecond latency and very high throughput are required. Nevertheless, the price per capacity ofsolid-state drives (SSDs) is relatively high, which has encouraged system architects to adoptdata reduction techniques, mainlydeduplication andcompression, in enterprise storage solutions. To provide higher reliability and performance, SSDs are typically grouped usingredundant array of independent disk (RAID) configurations. Data reduction on top of RAID arrays, however, adds I/O overheads and also complicates the I/O patterns redirected to the underlying backend SSDs, which invalidates the best-practice configurations used in AFS. Unfortunately, existing works on the performance of data reduction do not consider its interaction and I/O overheads with other enterprise storage components including SSD arrays and RAID controllers. In this paper, using a real setup with enterprise-grade components and based on the open-source data reduction module RedHat VDO, we reveal novel observations on the performance gap between the state-of-the-art and the optimal all-flash storage stack with integrated data reduction. We therefore explore the I/O patterns at the storage entry point and compare them with those at the disk subsystem. Our analysis shows a significant amount of I/O overheads for guaranteeing consistency and avoiding data loss through data journaling, frequent small-sized metadata updates, and duplicate content verification. We accompany these observations with cross-layer optimizations to enhance the performance of AFS, which range from deriving new optimal hardware RAID configurations up to introducing changes to the enterprise storage stack. By analyzing the characteristics of I/O types and their overheads, we propose three techniques: (a) application-aware lazy persistence, (b) a fast, read-only I/O cache for duplicate verification, and (c) disaggregation of block maps and data by offloading block maps to a very fast persistent memory device. By consolidating all proposed optimizations and implementing them in an enterprise AFS, we show 1.3× to 12.5× speedup over the baseline AFS with 90% data reduction, and from 7.8× up to 57× performance/cost improvement over an optimized AFS (with no data reduction) running applications ranging from 100% read-only to 100% write-only accesses.

Ultra Low-Latency Backscatter for Fast-Moving Location Tracking

This paper explores building an ultra-low latency and high-accuracy location tracking solution using battery-free tags. While there is rich prior work on location tracking with battery-free RFID tags and backscatter devices, these systems typically face tradeoffs with accuracy, power consumption, and latency. Such limitations make these existing solutions unsuitable for emerging applications like industrial augmented reality which requires tracking fast-moving machinery; monitoring indoor sports activities that require real-time tracking of fast-moving objects with high precision and under stringent latency constraints. We propose and demonstrate FastLoc, a precision tracking system that locates tiny, battery-free analog backscatter tags at sub-millisecond latency and sub-centimeter accuracy. FastLoc is a hybrid system that simultaneously uses RF and optical signals to track tiny tags that can be attached to everyday objects. FastLoc leverages the RF channel responses from tags for estimating the coarse region where the tags may be located. It simultaneously uses the sensed optical information modulated on the backscatter signals to enable fine-grained location estimation within the coarse region. To achieve this, we design and fabricate a custom analog tag that consumes less than 150 uW and instantaneously converts incident optical signals to one-shot wideband harmonic RF responses at nanosecond latency. We then develop a static high-density distributed-frequency structured light pattern that can localize tags in the area of interest at a sub-centimeter accuracy and microsecond-scale latency. A detailed experimental evaluation of FastLoc shows a median accuracy of 0.7 cm in tag localization with a 0.51 ms effective localization latency.

Six Critical Challenges for 6G Wireless Systems: A Summary and Some Solutions

Numerous articles are now appearing on 6G wireless systems, covering different aspects, ranging from vision, architecture, and applications to technology breakthroughs. With cellular systems in mind, this article presents six critical, yet fundamental, challenges that must be overcome before the development and deployment of 6G systems. These include opening the subterahertz spectrum for increased bandwidths and the ability to utilize these bandwidths, pushing the limits of semiconductor technologies for operation within these bands, developing transceiver designs and architectures to realize the high peak data rates, and achieving submillisecond latencies at the network level to achieve 6G key performance indicators (KPIs). Additionally, since 6G systems will not be introduced in a greenfield environment, backward compatibility with existing systems is discussed. Where possible, we present practical solutions to realize the identified challenges.

GPU-Accelerated Partially Linear Multiuser Detection for 5G and Beyond URLLC Systems

In this feasibility study, we have implemented a recently proposed partially linear multiuser detection algorithm in reproducing kernel Hilbert spaces (RKHSs) on a GPU-accelerated platform. Partially linear multiuser detection, which combines the robustness of linear detection with the power of nonlinear methods, has been proposed for a massive connectivity scenario with the non-orthogonal multiple access (NOMA). This is a promising approach, but detecting payloads within a received orthogonal frequency-division multiplexing (OFDM) radio frame requires the execution of a large number of inner product operations, which are the main computational burden of the algorithm. Although inner-product operations consist of simple kernel evaluations, their vast number poses a challenge in ultra-low latency (ULL) applications, because the time needed for computing the inner products might exceed the sub-millisecond latency requirement. To address this problem, this study demonstrates the acceleration of the inner-product operations through massive parallelization. The result is a GPU-accelerated real-time OFDM receiver that enables sub-millisecond latency detection to meet the requirements of 5th generation (5G) and beyond ultra-reliable and low latency communications (URLLC) systems. Moreover, the parallelization and acceleration techniques explored and demonstrated in this study can be extended to many other signal processing algorithms in Hilbert spaces, such as those based on projection onto convex sets (POCS) and adaptive projected subgradient method (APSM) algorithms. Experimental results and comparisons with the state-of-art confirm the effectiveness of our techniques.

Design of Edge Computing for 5G-Enabled Tactile Internet-Based Industrial Applications

5G-enabled Tactile Internet (TI) will rely on sensors, actuators, and ultra-reliable and ultralow latency communication for providing haptic human-machine interaction services. However, stringent requirements in terms of high reliability and sub-millisecond latency pose many daunting challenges that have limited the development of Tl services. In this article, we shed light on the design of edge computing for Tl industrial applications. We discuss major industrial use cases that will be transformed with the development of TI. We thoroughly present the benefits and challenges of the processing of Tl computation-in-tensive tasks in edge computing servers. Moreover, we highlight the current designs of edge resource allocation algorithms for tasks offloading and haptic and non-haptic feedback delivery in Tl industrial applications. Furthermore, we thoroughly discuss the current missing building blocks in edge computing to unleash TI industrial applications. Finally, open challenges that emerge when designing edge computing for supporting Tl industrial applications are discussed.

An overview on integrated localization and communication towards 6G

while the fifth generation (5G) cellular system is being deployed worldwide, researchers have started the investigation of the sixth generation (6G) mobile communication networks. Although the essential requirements and key usage scenarios of 6G are yet to be defined, it is believed that 6G should be able to provide intelligent and ubiquitous wireless connectivity with Terabits per second (Tbps) data rate and sub-millisecond (sub-ms) latency over three-dimensional (3D) network coverage. To achieve such goals, acquiring accurate location information of the mobile terminals is becoming extremely useful, not only for location-based services but also for improving wireless communication performance in various ways such as channel estimation, beam alignment, medium access control, routing, and network optimization. On the other hand, the advancement of communication technologies also brings new opportunities to greatly improve the localization performance, as exemplified by the anticipated centimeter-level localization accuracy in 6G by extremely large-scale multiple-input multiple-output (MIMO) and millimeter wave (mmWave) technologies. In this regard, a unified study on integrated localization and communication (ILAC) is necessary to unlock the full potential of wireless networks for dual purposes. While there are extensive studies on wireless localization or communications separately, the research on ILAC is still in its infancy. Therefore, this article aims to give a tutorial overview on ILAC towards 6G wireless networks. After a holistic survey on wireless localization basics, we present the state-of-the-art results on how wireless localization and communication inter-play with each other in various network layers, together with the main architectures and techniques for localization and communication co-design in current two-dimensional (2D) and future 3D networks with aerial-ground integration. Finally, we outline some promising future research directions for ILAC.

Instantaneous microwave-photonic spatial-spectral channelization via k-space imaging

The ability to both spatially and spectrally demultiplex wireless transmitters enables communication networks with higher spectral and energy efficiency. In practice, demultiplexing requires sub-millisecond latency to map the dynamics of the user space in real-time. Here, we present a system architecture, referred to as k-space imaging, which channelizes the radio frequency signals both spatially and spectrally through optical beamforming, where the latency is limited only by the speed of light traversing the optical components of the receiver. In this architecture, a phased antenna array samples radio signals, which are then coupled into electro-optic modulators (EOM) that coherently up-convert these signals to the optical domain, preserving their relative phases. The received signals, now optical sidebands, are transmitted in optical fibers of varying path lengths, which act as true time delays that yield frequency-dependent optical phases. The output facets of the optical fibers form a two-dimensional optical phased array in an arrangement preserving the phases generated by the angle of arrival (AoA) and the time-delay phases. Directing the beams emanating from the fibers through an optical lens produces a two-dimensional Fourier transform of the optical field at the fiber array. Accordingly, the optical beam formed at the back focal plane of the lens is steered based upon the phases, providing the angle of arrival and instantaneous frequency measurement (IFM), with latency determined by the speed of light over the optical path length. We present a numerical evaluation and experimental demonstration of this passive AoA- and frequency-detection capability.

Dynamic Buffer Sizing and Pacing as Enablers of 5G Low-Latency Services

3GPP standards organization is performing an impressive effort trying to reach sub-millisecond latencies for 5G. However, such efforts may become fruitless if exogenously generated delays at transport layer are not considered. Nowadays, Radio Access Networks (RANs) are deployed with large buffers to achieve full utilization and avoid squandering wireless resources. Unfortunately, and since the data path’s bottleneck resides on the radio link, RAN’s buffers are bloated by TCP’s congestion control algorithm. Thus, a flow with low-latency requirements that encounters a bloated buffer, suffers from inevitable large sojourn times associated with the buffer depletion time, severely downgrading its Quality of Service (QoS). This paper presents different solutions for efficiently multiplexing distinct traffic patterns that share buffers on the 5G stack. Bufferbloat is extensively studied within the actual 5G QoS scenario, which presents multiple challenges inherited from the dynamic radio link nature and the presence of multiple queues at different entities. We propose and extensively emulate different algorithms in order to avoid the exogenous delay caused by the bufferbloat phenomena. We use real cellular network traces with realistic delay-sensitive and background traffic patterns in different scenarios. The outcome presents valuable insights in the algorithms that will enable low-latency services to be delivered through the 5G network stack satisfying restrictive envisioned constraints.

Elmo: Source Routed Multicast for Public Clouds

We present Elmo, a system that addresses the multicast scalability problem in multi-tenant datacenters. Modern cloud applications frequently exhibit one-to-many communication patterns and, at the same time, require sub-millisecond latencies and high throughput. IP multicast can achieve these requirements but has control- and data-plane scalability limitations that make it challenging to offer it as a service for hundreds of thousands of tenants , typical of cloud environments. Tenants, therefore, must rely on unicast-based approaches ( e.g. , application-layer or overlay-based) to support multicast in their applications, imposing bandwidth and end-host CPU overheads, with higher and unpredictable latencies. Elmo scales network multicast by taking advantage of emerging programmable switches and the unique characteristics of data-center networks; specifically, the hypervisor switches, symmetric topology, and short paths in a datacenter. Elmo encodes multicast group information inside packets themselves, reducing the need to store the same information in network switches. In a three-tier data-center topology with 27,000 hosts, Elmo supports a million multicast groups using an average packet-header size of 114 bytes (max. 325 bytes), requiring as few as 1,100 multicast group-table entries on average in leaf switches, and having a traffic overhead as low as 5% over ideal multicast.

Low-Latency VLC System with Fresnel Receiver for I2V ITS Applications

This work presents a characterization of a low-cost, low-latency Visible Light Communication (VLC) prototype for infrastructure-to-vehicle (I2V) communication for future Intelligent Transportation Systems (ITS). The system consists of a regular traffic light as a transmitter (the red light is modulated with the information), and a photodetector as a receiver. The latter is equipped with low-cost Fresnel lenses as condensers, namely, 1 ′ ′ Fresnel and 2 ′ ′ Fresnel, to increase the optical gain of the system at the receiver. The system is capable of Active Decode and Relay (ADR) of information to further incoming units. The experimental characterization of amplitude and Packet Error Rate (PER) for the proposed system has been performed for distances up to 50 m. The results show that by incorporating the 2 ′ ′ Fresnel lens in the photodetector, an error free ( PER ≤ 10 − 5 ) I2V communication is established up to 50 m. Furthermore, the prototype can be used for both broadcast and beaconing transmission modes. This low-cost VLC-based system could offer sub-millisecond latency in the full ADR process for distances up to 36 m, which makes it suitable for integration in Cellular-V2X (C-V2X) and 5G platforms.

MGFAST: A New Generation of Copper Broadband Access

The ITU-T project G.mgfast defines the MGFAST technology providing aggregate bit rates of up to 10 Gb/s over twisted pair and coaxial wiring. The first volume of the Recommendation is to be completed in January 2020. Multi-gigabitper- second access to the customer premises is achieved by substantial expansion of frequency spectrum, advanced coding, and use of full-duplex transmission. In addition, MGFAST can provide sub-millisecond latency, enhanced support of multiple quality of service grades, and point-to-multipoint operation.

Providing Low Latency Guarantees for Slicing-Ready 5G Systems via Two-Level MAC Scheduling

5G comes with the promise of sub-millisecond latency, which is critical for realizing an array of emerging URLLC services, including industrial, entertainment, telemedicine, automotive, and tactile Internet applications. At the same time, slicing-ready 5G networks face the challenge of accommodating other heterogeneous coexisting services with different and potentially conflicting requirements. Providing latency and reliability guarantees to URLLC service slices is thus not trivial. We identify transmission scheduling at the RAN level as a significant contributor to end-toend latency when considering network slicing. In this direction, we propose a two-level MAC scheduling framework that can effectively handle uplink and downlink transmissions of network slices of different characteristics over a shared RAN, applying different per-slice scheduling policies, and focusing on reducing latency for URLLC services. Our scheme offers the necessary flexibility to dynamically manage radio resources to meet the stringent latency and reliability requirements of URLLC, as demonstrated by our simulation results.

On the Need of Joint Bandwidth and NFV Resource Orchestration: A Realistic 5G Access Network Use Case

5G envisages a “hyper-connected society,” where an enormous number of devices are inter-connected anywhere and at any time. Cloud-enabled radio access networks (RANs) where intelligence is placed in conjunction with the radio heads at the proximity of end users are a promising solution to fulfill the 5G expectations of sub-millisecond latency, huge traffic volumes, and higher data rates. Network functions virtualization (NFV) and software-defined networking (SDN) developments enable end users to access advanced features, such as configurability, automation, scalability, improved resource utilization, and multi tenancy over the cloud-enabled RANs. Management and orchestration techniques are the ultimate factor that determines the effectiveness of the novel SDN/NFV features being introduced. Our focus in this letter is the resource allocation in a realistic cloud-enabled RAN, taking into account the dynamics of ~100,000 persons movement in a crowded event, i.e., a football match. The proposed solution jointly orchestrates NFV and bandwidth resources, as one resource affects the other. Simulation results clearly verify the benefits of the proposed solution over traditional disjoint schemes.