Dell Products Inc.

Hagay Dagan

Gil Ratsaby

Ron Arnan

Adnan Sahin

Paulo Abelha Ferreira

Pablo Nascimento Da Silva

Adriana Bechara Prado

Krzysztof Misan

A method comprising: receiving a first workload data set, the first workload data set specifying a cache hit outcome distribution that is associated with a plurality of input-output (I/O) operations; identifying a plurality of workload portions of the first workload data set, each of the workload portions identifying: (i) a rate of a cache hit outcome that is associated with a respective I/O operation, and (ii) a data size that is associated with the respective I/O operations; generating a plurality of initial vectors, each of the initial vectors being generated based on a different one of the plurality of workload portions, each of the initial vectors being generated by a different sub-network of a correlation neural network; generating a context vector based on the plurality of initial vectors; processing the context vector with a decoder to generate a plurality of data points in a response curve of a storage system.

Response time estimation using sequence-to-sequence deep neural networks

One or more aspects of the present disclosure relate to predicting one or more performance metrics of a storage array configuration. In embodiments, a storage array configuration can be generated based on one or more array-configuration related parameters. In addition, IO workload metrics corresponding to an IO workload received by an array corresponding to the storage array configuration can be recorded. Further, at least one response time (RT) prediction can be generated based on the IO workload metrics. Accordingly, another storage array configuration can be generated based on the RT predictions and according to performance requirements defined by a service level objective (SLO).

Predicting storage array response times

Telemetry data gathered from active deployed SAN nodes is used to create a machine learning model that predicts storage system performance, e.g. in terms of response time. The telemetry data may be filtered to remove outlier values and less relevant information before creating the training dataset. Engineered features may be created that include types of data that are not present in the telemetry data. For example, data types from the telemetry data may be combined to create engineered features that are more relevant than the individual data types. The engineered features are included in the training dataset. The machine learning model may be used to test possible configurations for a planned SAN node based on expected workload and performance requirements. Outputted data may include satisfactory configurations for a planned storage system.

Using telemetry data from different storage systems to predict response time

A complex object generator is implemented, for example, as an integrated development environment. The complex object generator includes a hierarchical object relationship data structure describing classes of objects, relationships between the classes of objects, and metrics associated with the classes of objects. The hierarchical object relationship data structure is parsed by parser to create a hierarchy of Java classes. A user interface uses the hierarchy of Java classes to constrain selection of objects and metrics during creation of the complex object. As input is received relative to selected objects and metrics, the complex object is incrementally built. By constraining object and metric selection using the hierarchy of Java classes, the complex object is guaranteed to be valid when built, thus reducing or eliminating the number of errors associated with building complex objects.

Growing community of inventors

Pardes Hanna, Israel

Hagay Dagan