Distribution of global namespace to achieve performance and capacity linear scaling in cluster filesystems
AssigneeEMC IP Holding Company LLC
InventorsMurthy V. Mamidi, Abhishek Rajimwale, George Mathew
Life of this patent
Estimating the term…
Abstract
The distribution of global namespace to achieve performance and capacity linear scaling in cluster filesystems. Specifically, the disclosure entails the distribution of files, tracked in namespace trees, among various cluster nodes while linking the namespace trees through a centralized access point, thus providing a global namespace abstraction for the distributed cluster.
Claims20
A system, comprising:
- a plurality of data deduplication services (DDS) comprising a first DDS operatively connected to a second DDS,
- wherein the first DDS is programmed to:
- receive a file operation comprising a filename associated with a file;
- perform a lookup on a local redirection cache using the filename;
- make a first determination, based on the lookup, that the local redirection cache excludes the filename;
- traverse, based on the first determination, a meta-namespace tree using the filename, to identify a leaf page;
- make a second determination, based on examining the leaf page, that the leaf page excludes a child namespace record;
- identify, based on the second determination, a child indirect namespace record in the leaf page instead;
- obtain a first collection partition identifier (CPID) from the child indirect namespace record;
- and
- issue a remote procedure call (RPC) comprising the filename and a second CPID, to the second DDS,
- wherein the second DDS is associated with the first CPID.
A method for processing file operations, comprising:
- receiving a first file operation comprising a first filename associated with a first file;
- performing a first lookup on a local redirection cache using the first filename;
- making a first determination, based on the first lookup, that the local redirection cache excludes the first filename;
- traversing, based on the first determination, a meta-namespace tree using the first filename, to identify a first leaf page;
- making a second determination, based on examining the first leaf page, that the first leaf page excludes a child namespace record;
- identifying, based on the second determination, a child indirect namespace record in the first leaf page instead;
- obtaining a first collection partition identifier (CPID) from the child indirect namespace record;
- and
- issuing a first remote procedure call (RPC) comprising the first filename and a second CPID, to a first remote node associated with the first CPID,
- wherein the first remote node processes the first RPC to execute the first file operation.
A non-transitory computer readable medium (CRM) comprising computer readable program code, which when executed by a computer processor, enables the computer processor to:
- receive a first file operation comprising a first filename associated with a first file;
- perform a first lookup on a local redirection cache using the first filename;
- make a first determination, based on the first lookup, that the local redirection cache excludes the first filename;
- traverse, based on the first determination, a meta-namespace tree using the first filename, to identify a first leaf page;
- make a second determination, based on examining the first leaf page, that the first leaf page excludes a child namespace record;
- identify, based on the second determination, a child indirect namespace record in the first leaf page instead;
- obtain a first collection partition identifier (CPID) from the child indirect namespace record;
- and
- issue a first remote procedure call (RPC) comprising the first filename and a second CPID, to a first remote node associated with the first CPID,
- wherein the first remote node processes the first RPC to execute the first file operation.
Dependencies are inferred from the claim text.