公开(公告)号：US20140156601A1

公开(公告)日：2014-06-05

申请号：US14085656

申请日：2013-11-20

Applicant: NetApp, Inc.

Inventor： John Strunk ,  Douglas Santry ,  Sriram Venketaraman ,  Vaijayanti Bharadwaj

IPC: G06F3/06 ,  G06F17/30

CPC classification number: G06F3/0641 ,  G06F3/0619 ,  G06F3/067 ,  G06F17/30088 ,  G06F17/30094

Abstract: Systems and methods for hierarchical reference counting via sibling trees are provided. The hierarchical data structure, together with its associated operations, can efficiently maintain reference counts and significantly reduce input/output (IO) operations compared to traditional techniques. The data structure presented here is applicable to any directed acyclic graph (DAG-type) structure where reference counts are used. Various embodiments of the present invention use a data structure to maintain a “sibling pointer” (pointing to the sibling node as a way to avoid reference count updates) and a “sibling count.” When nodes in the tree diverge, the sibling pointer and sibling count are updated as opposed to directly manipulating the reference counts of the children of the diverging nodes. Various other embodiments can use additional entries or fields that allow for improved efficiency and advantages.

Abstract translation: 提供了通过兄弟树分级引用计数的系统和方法。 与传统技术相比，层次数据结构及其相关操作可以有效地保持参考计数并显着降低输入/输出（IO）操作。 这里呈现的数据结构适用于使用参考计数的任何有向无环图（DAG型）结构。 本发明的各种实施例使用数据结构来保持“同级指针”（指向同级节点作为避免参考计数更新的方式）和“兄弟数量”。当树中的节点发散时，兄弟节点指针和 兄弟姐妹计数被更新，而不是直接操纵分歧节点的孩子的参考计数。 各种其他实施例可以使用允许提高效率和优点的附加条目或字段。

公开(公告)号：US10798159B2

公开(公告)日：2020-10-06

申请号：US15659930

申请日：2017-07-26

Applicant: NetApp, Inc.

Inventor： Lars Eggert ,  Douglas Santry

IPC: H04L29/08 ,  H04L29/06 ,  H04L29/14 ,  G06F3/06

Abstract: Methods, non-transitory computer readable media, and computing devices that managing network traffic in a storage system include parse one or more packets received from a client to identify header data, and wherein the parsing comprises identifying when the one or more packets is a read request or a write request. A server is identified based on the parsed one or more packets. One or more rules are modified in the parsed one or more packets to allow data exchange between the client and the identified server. The parsed one or more packets are transmitted with the modified one or more rules to the identified server, wherein data exchange is initiated between the client and the identified server based on the parsed one or more packets with the modified one or more rules.

公开(公告)号：US20190037013A1

公开(公告)日：2019-01-31

申请号：US15659930

申请日：2017-07-26

Applicant: NetApp, Inc.

Inventor： Lars Eggert ,  Douglas Santry

IPC: H04L29/08

Abstract: Methods, non-transitory computer readable media, and computing devices that managing network traffic in a storage system include parse one or more packets received from a client to identify header data, and wherein the parsing comprises identifying when the one or more packets is a read request or a write request. A server is identified based on the parsed one or more packets. One or more rules are modified in the parsed one or more packets to allow data exchange between the client and the identified server. The parsed one or more packets are transmitted with the modified one or more rules to the identified server, wherein data exchange is initiated between the client and the identified server based on the parsed one or more packets with the modified one or more rules.

公开(公告)号：US09274716B2

公开(公告)日：2016-03-01

申请号：US14085656

申请日：2013-11-20

Applicant: NetApp, Inc.

Inventor： John Strunk ,  Douglas Santry ,  Sriram Venketaraman ,  Vaijayanti Bharadwaj

IPC: G06F17/30 ,  G06F3/06

CPC classification number: G06F3/0641 ,  G06F3/0619 ,  G06F3/067 ,  G06F17/30088 ,  G06F17/30094

Abstract: Systems and methods for hierarchical reference counting via sibling trees are provided. The hierarchical data structure, together with its associated operations, can efficiently maintain reference counts and significantly reduce input/output (IO) operations compared to traditional techniques. The data structure presented here is applicable to any directed acyclic graph (DAG-type) structure where reference counts are used. Various embodiments of the present invention use a data structure to maintain a “sibling pointer” (pointing to the sibling node as a way to avoid reference count updates) and a “sibling count.” When nodes in the tree diverge, the sibling pointer and sibling count are updated as opposed to directly manipulating the reference counts of the children of the diverging nodes. Various other embodiments can use additional entries or fields that allow for improved efficiency and advantages.

Abstract translation: 提供了通过兄弟树分级引用计数的系统和方法。 与传统技术相比，层次数据结构及其相关操作可以有效地保持参考计数并显着降低输入/输出（IO）操作。 这里呈现的数据结构适用于使用参考计数的任何有向无环图（DAG型）结构。 本发明的各种实施例使用数据结构来保持“同级指针”（指向同级节点作为避免参考计数更新的方式）和“兄弟数量”。当树中的节点发散时，兄弟节点指针和 兄弟姐妹计数被更新，而不是直接操纵分歧节点的孩子的参考计数。 各种其他实施例可以使用允许提高效率和优点的附加条目或字段。