Microsoft Technology Licensing, LLC

Netapp, Inc.

Thomas M Talpey

Gregory Robert Kramer

David Matthew Kruse

James T Pinkerton

James F Lentini

Peter Shah

Jason L Goldschmidt

Alexandros Batsakis

Richard John Black

Antony Ian Taylor Rowstron

Hitesh Ballani

Thomas Karagiannis

Gregory O'Shea

Eno Thereska

Arkady Kanevsky

Gregory O'SHEA

The embodiments described herein generally relate to a protocol for implementing data operations, e.g., a version of SMB, atop RDMA transports. In embodiments, systems and methods use the protocol definition, which specifies new messages for negotiating an RDMA connection and for transferring SMB2 data using the negotiated communication. A new protocol message may include new header information to determine message size, number of messages, and other information for sending the SMB2 data over RDMA. The header information is used to accommodate differences in message size requirements between RDMA and SMB2. The SMB Direct protocol allows SMB2 data to be fragmented into multiple individual RDMA messages that a receiver may then logically concatenate into a single SMB2 request or SMB2 response. The SMB Direct protocol also may allow SMB2 to transfer application data via efficient RDMA direct placement and to signal the application data's availability when the transfer is complete.

Transporting operations of arbitrary size over remote direct memory access

The moderation of event notifications from a network interface card. The network interface card has multiple completion queues that queue of completed work. The moderation batches up this completed work such that potentially multiple work requests are aggregated into a single event notification. This moderation reduces processing overhead since it spreads the overhead associated with a single interrupt to multiple event notifications The decision on moderation may be performed per connection, or even per constituent queue of the connection. The principles herein allow moderation to reduce overhead without slowing network throughput.

Moderated completion signaling

Controlling data storage input/output requests is described, for example, to apply a policy to an end-to-end flow of data input/output requests between at least one computing entity and at least one store. In various examples a plurality of queues are configured at one or more stages of the end-to-end flow and controlled to adhere to a policy. In examples, each stage has a control interface enabling it to receive and execute control instructions from a controller which may be centralized or distributed. For example, the control instructions comprise queuing rules and/or queue configurations. In various examples queues and queuing rules are dynamically created and revised according to feedback about any of: flow behavior, changes in policy, changes in infrastructure or other factors. In examples, high level identifiers of the flow endpoints are resolved, on a per stage basis, to low level identifiers suitable for use by the stage.

Systems and methods based on policy criteria for controlling the flow of data storage input/output requests between endpoints

Embodiments described herein are directed to extending remote direct memory access (RDMA) semantics to enable implementation in a local storage system and to providing a management interface for initializing a local data store. A computer system extends RDMA semantics to provide local storage access using RDMA, where extending the RDMA semantics includes the following: mapping RDMA verbs of an RDMA verbs interface to a local data store and altering RDMA ordering semantics to allow out-of-order processing and/or out-of-order completions. The computer system also accesses various portions of the local data store using the extended RDMA semantics.

Mapping RDMA semantics to high speed storage

A flow based reply cache of a storage system is illustratively organized into one or more microcaches, each having a plurality of reply cache entries. Each microcache is maintained by a protocol server executing on the storage system and is allocated on a per client basis. To that end, each client is identified by a client connection or logical 'data flow' and is allocated its own microcache and associated entries, as needed. As a result, each microcache of the reply cache may be used to identify a logical stream of client requests associated with a data flow, as well as to isolate that client stream from other client streams and associated data flows used to deliver other requests served by the system. The use of microcaches thus provides a level of granularity that enables each client to have its own pool of reply cache entries that is not shared with other clients, thereby obviating starvation of entries allocated to the client in the reply cache.

Flow based reply cache

Scheduling operations such as asynchronous file system operations in a network storage system is accomplished by applying a bid-price online auction methodology, in which bid (willingness-to-pay) values and price (cost) values are dynamically set by storage clients and a storage server, respectively, based on utilization of computing resources. The system provides a framework for adaptively scheduling asynchronous file system operations, managing multiple key resources of the distributed file system, including network bandwidth, server I/O, server CPU, and client and server memory utilization. The system can accelerate, defer, or cancel asynchronous requests to improve application-perceived performance. Congestion pricing via online auctions can be employed to coordinate the use of system resources by clients, so clients can detect shortages and adapt their resource usage.

Adaptive scheduling of storage operations based on utilization of a multiple client and server resources in a distributed network storage system

A system and method determines when the entries of a reply cache, organized into microcaches each of which is allocated to a client connection, may be retired or released, thereby freeing up memory structures. A plurality of connection statistics are defined and tracked for each microcache and for the entries of the microcache. The connection statistics indicate the value of the microcache and its entries to the client. The connection statistics include a measure of the time since the last idempotent or non-idempotent request (TOLR) was received, and a count of the number of idempotent requests that have been received since the last non-idempotent request (RISLR). A microcache with a TOLR time and a RISLR count that exceed respective thresholds may be expired and removed from the reply cache.

System and method for managing a flow based reply cache

A computer-implemented method and apparatus for enumerating program code dependencies is provided. According to an embodiment of the invention, a user selects or otherwise indicates a code base to be processed and analyzed for the purpose of enumerating those program code modules required by the code base in order to properly compile and execute the code base. Next, the code base is processed in accordance with instructions designed to mirror those of a compiler for compiling the code base. Accordingly, external program code dependencies are identified using compile time compilation parameters, as well as extensive programming language analysis.

Growing community of inventors

Stow, MA, United States of America

Thomas M Talpey