Abstract: A system for improving task scheduling on a cloud data platform is provided. A task to be executed using resources of a computing cluster is received. A task execution plan is generated and information about data to be used for the ask is accessed. Resource requirements for executing the task are predicted by applying machine learning to the task execution plan and the information about the data. Assignment data is generated to execute the task on the resources by applying machine learning information about a current state of the resources and predicted resource requirements.

Abstract: Using an attributes model of a time series forecasting model, determine a set of features based on time series data, the set of features including periodic components. The time series data may be divided into a set of segments. For each segment of the set of segments, a weight may be assigned using an age of the segment, resulting in a set of weighted segments of time series data. Using a trend detection model of the time series forecasting model, trend data from the set of weighted segments of time series data may be determined. A time series forecast may be generated by combining the set of features and the trend data.

Abstract: A system for improving task scheduling on a cloud data platform is provided. A task is received, from a user of a cloud data platform, for execution on a dataset of a cloud data platform using a plurality of resources. A task graph is generated, and metadata related to the dataset is accessed for use in execution of the task. A predicted resource profile is generated by applying a first machine learning scheme to the task graph and the metadata of the dataset. Assignment data is generated to execute processes of the task on the plurality of resources. The assignment data generated by applying a second machine learning scheme to current state data of a current computational state of the plurality of resources and the predicted resource profile generated by the first machine learning scheme.

Abstract: Embodiments of the present disclosure may provide a database optimization system that can generate computational values through a database compiler and assignment data for execution of a query by a plurality of nodes of a database system. The computational values and assignment data can be generated by one or more machine learning schemes. The machine learning schemes can be trained on previous computational values and previous assignment data.

Abstract: A processing device may analyze a set of time series data using a time series forecasting model comprising an attributes model and a trend detection model. The attributes model may comprise a modified gradient boosting decision tree (GBDT) based algorithm. Analyzing the set of time series data comprises determining a set of features of the set of time series data, the set of features including periodic components as well as arbitrary components. A trend of the set of time series data may be determined using the trend detection model and the set of features and the trend may be combined to generate a time series forecast.

Abstract: This invention provides a new mechanism to load-balance traffic using only a SDN switch with high TCAM space efficiency, avoidance of frequent updates, robustness against accidental or malicious traffic overload, and balancing with respect to any load metric provided said metric is monotonically increasing with traffic rates. Layer for load-balancing logic is folded into the invention by the introduction of L4 matches and return flow-pinning.

Abstract: This invention provides a new mechanism to provide performance isolation between the different port-and-direction pairs of a software switch. This is accomplished by mapping each port-pair and direction to its own Operating System process. This improves performance, fault, and rule-space isolation between the ports of a software-based network switch on general purpose CPUs. This invention makes it possible to use standard OS mechanisms and commands to control the per-port isolation of network packet forwarding on a software switch.

Abstract: In a software-defined network (SDN) multiple servers and switches (the devices) connect into a (sub-)network under a (logically) centralized controller. As each device boots up it may have physically network connectivity but logically it is not connected to the SDN. In order to establish a logical connection the device must receive a network address assignment. There are legacy protocols for this purpose, e.g., DHCP that could be and have been adapted for use in software-defined networks. This invention specifies a different, enumeration-based method for address assignment in hierarchically-organized SDNs, which exploits the topological properties of the SDN for the purpose of device enumeration. This invention provides both Layer 2 and Layer 3 addresses for each device of the SDN.

Abstract: This present invention describes the use of a uni-directional radio channel to be used for communication from a Controller to a remote Router if the wired Internet connection that connects a Controller to Router becomes unavailable in the direction from the Controller to the Router. The invention provides a slow but widely available uni-directional long-range radio based backup channel that can be used to remotely fix a router misconfiguration that may have caused the disconnection, most likely by switching said router into a safe default mode.

Abstract: Techniques are provided for detecting computer fraud. The techniques include obtaining a text version of a candidate destination and a graphical rendering of the candidate destination, comparing the text version of the candidate destination and the graphical rendering of the candidate destination with a corresponding text version of a stored destination and a corresponding graphical rendering of the stored destination, and generating a fraud warning if the graphical rendering of the candidate destination is substantially similar to the graphical rendering of the stored destination while the text version of the candidate destination differs substantially from the corresponding text version of the stored destination.

Abstract: Techniques are provided for computer network security. The techniques include obtaining operational data for at least a first networked application; obtaining enterprise data for at least a second networked application; correlating the operational data with the enterprise data to obtain correlated data; and using the correlated data to improve security of the computer network.

Abstract: Techniques are provided for automatically generating a database of destinations to be protected against computer fraud. The techniques include generating at least one category of destinations to be protected, and retrieving (i) at least one list of destinations belonging to said at least one category and (ii) contents specified at each uniform resource locator associated with a destination identified in the at least one list of destinations belonging to said at least one category.

Abstract: A computer implemented method of virtual machine migration with filtered network connectivity includes enforcing network security and routing at a hypervisor layer which is independent of guest operating systems via dynamic updating of routing controls initiated by a migration of said virtual machine from a first device to a second device.

Abstract: A distributed network system includes at least one server having memory and one or more processors. The server performs the following operations in connection with propagating object updates to respective client devices: receiving a first object update message from an application server, the first object update message including a first object identifier and a first object version; updating an object update record that includes the first object identifier using the first object version, wherein the object update record includes a first client identifier; propagating the first object version from the object update record to a client-object registration record that includes the first client identifier; and sending a second object update message to a first client device associated with the first client identifier, wherein the second object update message includes the first object identifier and the first object version.

Abstract: A distributed network system includes a distributed server system and a plurality of client devices that are communicatively coupled to the distributed server system. The distributed server system includes multiple servers, each server including a set of client devices registered with the server. Through client re-registration, a client device can move its registration from one server to another server. Through server-side data replication, an object registration request submitted by a client device to a first server is forwarded to a second server without having the client device submitting the same request to the second server.

Abstract: A computer implemented method of virtual machine migration with filtered network connectivity includes enforcing network security and routing at a hypervisor layer which is independent of guest operating systems via dynamic updating of routing controls initiated by a migration of said virtual machine from a first device to a second device.

Abstract: A method (and system) which provides virtual machine migration with filtered network connectivity and control of network security of a virtual machine by enforcing network security and routing at a hypervisor layer at which the virtual machine partition is executed, and which is independent of guest operating systems.

Abstract: Techniques are provided for detecting computer fraud. The techniques include obtaining a text version of a candidate destination and a graphical rendering of the candidate destination, comparing the text version of the candidate destination and the graphical rendering of the candidate destination with a corresponding text version of a stored destination and a corresponding graphical rendering of the stored destination, and generating a fraud warning if the graphical rendering of the candidate destination is substantially similar to the graphical rendering of the stored destination while the text version of the candidate destination differs substantially from the corresponding text version of the stored destination.

Abstract: A voice-over-internet protocol (VoIP) caller verification method includes generating a challenge to be sent to a VoIP caller which needs a response from the VoIP caller prior to accepting a VoIP call, wherein the challenge includes a test pattern crafted to create a greater load at a spammer VoIP caller as compared to a human VoIP caller. A response or lack thereof is determined from the VoIP caller, and the response is analyzed relative to an acceptable response to determine whether the VoIP caller is acceptable to communicate with.

Abstract: A proactive problem resolution system, method of proactive problem resolution and program product therefor. User sensors extract data from user interaction with a computer terminal and pass extracted data to a sniffer agent. The sniffer agent checks for an indication of a user problem by comparing user behavior data against behavior data from previously encountered problems. When the sniffer finds a match, the computer terminal user may be asked if assistance is needed or the user may be automatically referred to the help desk. If a solution already exists for a problem that corresponds to the user behavior, that solution(s) is(are) presented at the computer terminal. Computer terminal users automatically receive problem support, even before it would otherwise be requested.