Sriram S Rao, Age 5313458 Briar Ct, Saratoga, CA 95070

7467282, Dec 16, 2008

Apr 5, 2005

11/099925

Sriram Rao - San Jose CA, US
John Edwards - Sunnyvale CA, US
Douglas P. Doucette - Freeland WA, US
Cheryl Thompson - Sunnyvale CA, US

Network Appliance, Inc. - Sunnyvale CA

G06F 12/00

711202, 711165, 711170, 711171, 711172, 711173, 711203, 711205, 711206, 711207

A file system migrates a traditional volume to a virtual volume without data copying. In an embodiment, a traditional volume index node is selected for migration. The traditional volume index node is converted to a virtual volume index node. In one embodiment, the virtual volume index node provides both physical address information and virtual address information.

7730277, Jun 1, 2010

Oct 25, 2004

10/972817

Ashish Prakash - Morrisville NC, US
John K. Edwards - Sunnyvale CA, US
Sriram Rao - San Jose CA, US

NetApp, Inc. - Sunnyvale CA

G06F 12/00
G06F 13/00
G06F 13/28
G06F 9/26
G06F 9/34
G06F 9/455
G06F 17/30

711170, 711171, 711172, 711173, 711159, 711202, 711203, 718 1, 707206

A multi-stage technique invalidates and replaces loadable physical volume block numbers (pvbns) stored in indirect blocks of a dual vbn (“flexible”) virtual volume (vvol) of a storage system to enable efficient image transfers and/or fragmentation handling of the flexible vvol. Each loadable pvbn of a pvbn/virtual vbn (vvbn) block pointer pair is converted into a special block pointer having a predefined reserved value that provides a temporary “pvbn_unknown” placeholder until replaced by a real (actual) pvbn. The technique further allows the storage system to serve data from the flexible vvol using the placeholders while the actual pvbns are computed, thereby eliminating latencies associated with completion of actual pvbn replacement for the pvbn_unknown placeholders.

7774327, Aug 10, 2010

Jan 31, 2006

11/345249

Sriram S. Rao - San Jose CA, US
Sandeep Yadav - Santa Clara CA, US
John Z. Zhuge - San Jose CA, US

Network Appliance, Inc. - Sunnyvale CA

G06F 17/30

707705, 707821

A storage server maintains an active file system and at least one read-only, persistent point-in-time image (RPPI) of the active file system. The booting time of the storage server is reduced by delaying the mounting of the RPPI(s) until the storage server receives a request to access a corresponding RPPI after the storage server has already booted.

7836020, Nov 16, 2010

Apr 3, 2006

11/397466

Sriram S. Rao - San Jose CA, US
Joydeep Sen Sarma - Sunnyvale CA, US
Alan L. Rowe - San Jose CA, US

Network Appliance, Inc. - Sunnyvale CA

G06F 7/00
G06F 17/00

707649, 707685

A system is provided to improve storage server performance. The system comprises a receiving component of a storage server to receive a partner server's file system metadata while the partner server operates in a normal mode. The partner server's file system metadata is stored in the storage servers metadata cache and is used to mount the partner server's file system when the first server transitions into a takeover mode of operation.

2003000, Jan 9, 2003

Jun 22, 2001

09/887810

Sriram Rao - Mountain View CA, US

H04L012/56

370/390000, 370/432000

A method and apparatus for reducing performance overhead of a media server sending packetized audio/video data to an end-player or user is provided. The main sources of performance overhead associated with a media server transmitting packetized audio/video data are (1) copying data from the user space buffer to the kernel buffer and (2) context switching from user level to kernel level. There are several techniques that can be used to address overhead. One technique involves creating an aggregate packet, which includes a policy, at the user level and transmitting this aggregate packet from the user level to the OS level with a system call. A second technique involves manipulating packets based on a policy in the OS level rather than in the user level. The manipulation of packets may include redirecting packets or splitting packets.

2013000, Jan 3, 2013

Jul 1, 2011

13/175782

Sriram Rao - San Jose CA, US
Adam Silberstein - Sunnyvale CA, US
Russell Sears - Berkeley CA, US
Azza Abouzeid - New Haven CT, US

Yahoo! Inc. - Sunnyvale CA

G06F 15/16

709201

Various methods and apparatuses are provided which may be implemented using one or more computing devices within a networked computing environment to support a computing grid having selective storage of shared data files within certain distributed the systems provided by dusters of computing devices. The selective storage may represent limited duplicative storage of a shared file.

2021028, Sep 16, 2021

Jun 2, 2021

17/337280

- Redmond WA, US
Sriram S. Rao - Saratoga CA, US
Christopher W. Douglas - Mountain View CA, US
Ashvin Agrawal - Cupertino CA, US
Avrilia Floratou - Mountain View CA, US
Ishai Menache - Redmond CA, US
Srikanth Kandula - Redmond WA, US
Mainak Ghosh - Foster City CA, US
Joseph Naor - Haifa, IL

G06F 12/0862
H04L 29/08
G06F 12/0868
G06F 12/0891

Methods, systems, apparatuses, and computer program products are provided for prefetching data. A workload analyzer may identify job characteristics for a plurality of previously executed jobs in a workload executing on a cluster of one or more compute resources. For each job, identified job characteristics may include identification of an input dataset and an input bandwidth characteristic for the input dataset. A future workload predictor may identify future jobs expected to execute on the cluster based at least on the identified job characteristics. A cache assignment determiner may determine a cache assignment that identifies a prefetch dataset for at least one of the future jobs. A network bandwidth allocator may determine a network bandwidth assignment for the prefetch dataset. A plan instructor may instruct a compute resource of the cluster to load data to a cache local to the cluster according to the cache assignment and the network bandwidth assignment.

2021009, Apr 1, 2021

Oct 22, 2019

16/660554

- Redmond WA, US
Sriram S. Rao - Saratoga CA, US
Christopher W. Douglas - Mountain View CA, US
Ashvin Agrawal - Cupertino CA, US
Avrilia Floratou - Mountain View CA, US
Ishai Menache - Redmond CA, US
Srikanth Kandula - Redmond WA, US
Mainak Ghosh - Foster City CA, US
Joseph Naor - Haifa, IL

G06F 12/0862
G06F 12/0891
G06F 12/0868
H04L 29/08

Methods, systems, apparatuses, and computer program products are provided for prefetching data. A workload analyzer may identify job characteristics for a plurality of previously executed jobs in a workload executing on a cluster of one or more compute resources. For each job, identified job characteristics may include identification of an input dataset and an input bandwidth characteristic for the input dataset. A future workload predictor may identify future jobs expected to execute on the cluster based at least on the identified job characteristics. A cache assignment determiner may determine a cache assignment that identifies a prefetch dataset for at least one of the future jobs. A network bandwidth allocator may determine a network bandwidth assignment for the prefetch dataset. A plan instructor may instruct a compute resource of the cluster to load data to a cache local to the cluster according to the cache assignment and the network bandwidth assignment.