We present NetCache, a new key-value store architecture that leverages the power and flexibility of new-generation programmable switches to handle queries on hot items and balance the load across storage nodes. NetCache provides high aggregate throughput and low latency even under highly-skewed and rapidly-changing workloads. The core of NetCache is a packet-processing pipeline that exploits the capabilities of modern programmable switch ASICs to efficiently detect, index, cache and serve hot key-value items in the switch data plane. Additionally, our solution guarantees cache coherence with minimal overhead. We implement a NetCache prototype on Barefoot Tofino switches and commodity servers and demonstrate that a single switch can process 2+ billion queries per second for 64K items with 16-byte keys and 128-byte values, while only consuming a small portion of its hardware resources. To the best of our knowledge, this is the first time that a sophisticated application-level functionality, such as in-network caching, has been shown to run at line rate on programmable switches. Furthermore, we show that NetCache improves the throughput by 3-10x and reduces the latency of up to 40% of queries by 50%, for high-performance, in-memory key-value stores.

https://dl.acm.org/doi/pdf/10.1145/3132747.3132764

NetCache: Balancing Key-Value Stores with Fast In-Network Caching

A network accessible node that facilitates management of a fleet of portable communication devices (EIR terminal), including portable data terminals and/or barcode readers, by directing performance of software upgrade and/or configuration update actions by one or more members of the fleet of EIR terminals.

Data collection system having eir terminal interface node

A cloud computing environment provides the ability to deploy a web application that has been developed using one of a plurality of application frameworks and is configured to execute within one of a plurality of runtime environments. The cloud computing environment receives the web application in a package compatible with the runtime environment (e.g., a WAR file to be launched in an application server, for example) and dynamically binds available services by appropriately inserting service provisioning data (e.g., service network address, login credentials, etc.) into the package. The cloud computing environment then packages an instance of the runtime environment, a start script and the package into a web application deployment package, which is then transmitted to an application (e.g., container virtual machine, etc.). The application container unpacks the web application deployment package, installs the runtime environment, loads the web application package into the runtime environment and starts the start script, thereby deploying the web application in the application container.

Cloud platform architecture

Management of data communication between a peripheral device and host computer system is provided. A peripheral device exposes to a host computer system multiple interfaces for data communication between the peripheral device and the host computer system. The multiple interfaces are exposed over a single physical interface between the peripheral device and the host computer system, for communicating data between the peripheral device and multiple applications executing on the host computer system. The multiple interfaces can include a data collection interface facilitating collection of data from the peripheral device by an application of the multiple applications executing on the host computer system.

Managing data communication between a peripheral device and a host

Methods managing data communication between a peripheral device and host computer system are provided. A physical interface for communicating data between a peripheral device and the plurality of applications executing on the host computer system is opened and controlled by a software module. A first virtual interface and a second virtual interface of the software module are exposed to an operating system of the host computer system, and the operating system exposes the first virtual interface and the second virtual interface to the first application and the second application. The first virtual interface is used for communicating data between the peripheral device and the first application through the physical interface, and the second virtual interface is used for communicating data between the peripheral device and the second application through the physical interface.

Device management using virtual interfaces cross-reference to related applications

An invention is disclosed managing the replicated and migration capable state for an enterprise Java bean (EJB) application. The invention includes executing a Java application on a server that includes an entity bean. In addition, a replicated state manager is executed that includes program instructions for managing an in-memory state of the Java application, and program instructions for replicating the in-memory state of the Java application to a replicated state server. The replicated state server can be a memory replicated state server, or a disk replicated state server. To facilitate application state management, embodiments of the present invention store states of the entity beans objects using state objects, which are updated in response to changes in the state of the application. Hence, the embodiments of the present invention define a logical separation between the application and the state objects.

Method and apparatus for managing replicated and migration capable session state for a java platform

Method, system, and programs for resolving data inconsistency in a distributed system having a plurality of replica instances. In response to a read request of at least one data element, a key-specific probability value for the at least one data element of the read request is retrieved. Based on the retrieved key-specific probability value and a system load of the distributed system, a self-healing probability of the at least one data element of the read request is computed. A self-healing transaction that includes the at least one data element of the read request is then generated based on the self-healing probability of the at least one data element. The generated self-healing transaction is further applied to at least some of the plurality of replica instances in the distributed system to resolve data inconsistency.

Method and system for resolving data inconsistency

An invention for migrating managed application state for a Java based application is disclosed. A first Java module is executed on afirst server. The first Java module includes a first entity bean and a first state object in communication with the first entity bean. The first state object stores a state of the first entity bean. The first state object is replicated to a state server or stored in-memory and made capable of migration by replicated state manager. Then, a second Java module is started on a second server as part of the module migration by replicated state manager. Then, a second Java module is started on asecond server as part of the module migration by a control module. The second Java module includes a second state object that is populated with state by either recovering from replica of first state object stored on the state server or by getting transferred state of an in-memory copy of first state object managed on first server.

Method and apparatus for migration of managed application state for a java based application

This paper presents a fully convolutional scene graph generation (FCSGG) model that detects objects and relations simultaneously. Most of the scene graph generation frameworks use a pre-trained two-stage object detector, like Faster R-CNN, and build scene graphs using bounding box features. Such pipeline usually has a large number of parameters and low inference speed. Unlike these approaches, FCSGG is a conceptually elegant and efficient bottom-up approach that encodes objects as bounding box center points, and relationships as 2D vector fields which are named as Relation Affinity Fields (RAFs). RAFs encode both semantic and spatial features, and explicitly represent the relationship between a pair of objects by the integral on a sub-region that points from subject to object. FCSGG only utilizes visual features and still generates strong results for scene graph generation. Comprehensive experiments on the Visual Genome dataset demonstrate the efficacy, efficiency, and generalizability of the proposed method. FCSGG achieves highly competitive results on recall and zeroshot recall with significantly reduced inference time.

/pdf/fully-convolutional-scene-graph-generation-5glswrwnn6.pdf

Fully Convolutional Scene Graph Generation

Method, system, and programs for balancing work load in a distributed system. A plurality of multi-dimensional load metrics are received from a plurality of resource units in the distributed system. Based on the received plurality of multi-dimensional load metrics and a global statistical load model, a load deviance for each resource unit is computed. The plurality of resource units in the distributed system are then ranked based on the load deviance of each resource unit. At least one load balancing action is further determined based on the ranked resource units and at least one load balancing policy.

Masood Mortazavi

Papers

Method and apparatus for managing replicated and migration capable session state for a java platform

Method and system for resolving data inconsistency

Method and apparatus for migration of managed application state for a java based application

Fully Convolutional Scene Graph Generation

Method and system for work load balancing