OneFS file system overview:
The EMC OneFS file system is a distributed file system that runs on a storage cluster. It combines three layers of traditional storage architecture namely, a File System, Volume Manager, and data protection.
It scales out because it relies on intelligent software, commodity hardware, and distributed architecture.
OneFS works with a cluster that consists of multiple nodes and starts out with as few as three nodes. In the clustrer, nodes could provide different ratios of throughput and Input/Output operations per second. OneFS combines these into a whole - RAM is grouped together into a single coherent cache, NVRAM is grouped together for high throughput writes and spindles and CPU are combined to increase throughput, capacity and IOPS
There are two types of networks associated with a cluster : internal and external. Intra-node communication in a cluster is performed using a proprietary unicast node to node protocol. IP over InfiniBand network is used with redundant switches.
Clients connect to the cluster using Ethernet connections, each node provides its own ports. File system protocols such as NFS, CIFS, HTTP, iSCSI, FTP and HDFS are supported.
The Operating system of OneFS is a BSD-based operating system that supports both windows and Linux/Unix based semantics such as hard links, delete-on-close, atomic rename, ACLs and extended attributes.
To communicate with the client, a variety of protocols as mentioned above are supported. The I/O subsystem is split in two halves - the top half is the initiator and the bottom half is the participant. Any node that the client connects to acts as the initiator.
Cluster operations involve checking and maintaining the health of the cluster. These jobs are run through a job engine and have priority associated. Jobs might include balancing free space in a cluster, scanning for viruses, reclaiming disk space, associating a path and its contents with a domain, rebuild and re-protect the file system, gathers the file system, scrubs disk for media level errors, and revert an entire snapshot to disk.
The granularity of the jobs ensures that OneFS performs adequately and appropriately for every impact interval in the customers environment.
The EMC OneFS file system is a distributed file system that runs on a storage cluster. It combines three layers of traditional storage architecture namely, a File System, Volume Manager, and data protection.
It scales out because it relies on intelligent software, commodity hardware, and distributed architecture.
OneFS works with a cluster that consists of multiple nodes and starts out with as few as three nodes. In the clustrer, nodes could provide different ratios of throughput and Input/Output operations per second. OneFS combines these into a whole - RAM is grouped together into a single coherent cache, NVRAM is grouped together for high throughput writes and spindles and CPU are combined to increase throughput, capacity and IOPS
There are two types of networks associated with a cluster : internal and external. Intra-node communication in a cluster is performed using a proprietary unicast node to node protocol. IP over InfiniBand network is used with redundant switches.
Clients connect to the cluster using Ethernet connections, each node provides its own ports. File system protocols such as NFS, CIFS, HTTP, iSCSI, FTP and HDFS are supported.
The Operating system of OneFS is a BSD-based operating system that supports both windows and Linux/Unix based semantics such as hard links, delete-on-close, atomic rename, ACLs and extended attributes.
To communicate with the client, a variety of protocols as mentioned above are supported. The I/O subsystem is split in two halves - the top half is the initiator and the bottom half is the participant. Any node that the client connects to acts as the initiator.
Cluster operations involve checking and maintaining the health of the cluster. These jobs are run through a job engine and have priority associated. Jobs might include balancing free space in a cluster, scanning for viruses, reclaiming disk space, associating a path and its contents with a domain, rebuild and re-protect the file system, gathers the file system, scrubs disk for media level errors, and revert an entire snapshot to disk.
The granularity of the jobs ensures that OneFS performs adequately and appropriately for every impact interval in the customers environment.
No comments:
Post a Comment