Storage networking appliance

In my internship I got chance to deal closely with netapp systems.There I find interesting feature call snapshot,which is very useful to administrators to restore systems when something goes wrong.

Benefits of storage appliances

1. Easy to administration- File systems grow automatically when new disk are added.Logical partitions can be expanded and shrunk dynamically.Can take snapshots anytime(read only view of WAFL volume).
2. Increase performance by reducing complexity at a reasonable cost.
3. Data availability-RAID is used to protect data against data loss occured by disk failures.Snapshot, SnapMirror, SnapRestore can used as instant recovery.
4. development and scalability.

Storage appliance architecture

The storage appliance provides block-level data access over a Fiber Channel SAN fabric using FCP and over an IP-based Ethernet network using iSCSI. File access protocols such as NFS, CIFS, HTTP, or FTP provide file-level access over an IP-based Ethernet network.WAFL(the storage appliance's file system) and RAID were designed together to avoid the performance problems that most file systems cause with RAID.WAFL makes RAID4 operate efficiently,WAFL always writes blocks to stripes that are near each other, eliminating long seeks on the parity disk.

WAFL(Write anywhere file layout)

1. WAFL is block based, using 4KB blocks with no fragments.
2. WAFL uses inodes to describe its files.
3. Directories are specially formatted files.

WAFL uses files to store meta data,which allows mata-data blocks to be written anywhere on disk.WAFL introduced a new building block known as flexible volumes.

Snapshot technology

This is a benefit of the WAFL write anywhere approach. A Snapshot copy is an online, read-only copy of the entire file system(it takes less than 1 second time).A Snapshot copy can be used as an online backup capability(to tape), allowing users to recover their own files.

(a) is a simplified diagram of the file system, with a root structure at the top pointing to disk blocks.
(b) Snapshot copy being created by duplicating the root structure.
(c) shows block C being updated, so that the file system points to new data in C' while the Snapshot copy still points to the original data in C.

Every directory in the file system contains a special subdirectory that allows users to access the Snapshot copy.And the .snapshot directories are "hidden" in the sense that they do not show up in directory listings.
At least once every 10 seconds WAFL generates an internal Snapshot copy called a consistency point.When the storage appliance boots, WAFL always uses the most recent consistency point on disk, which means that even after power loss or system failure there is no need for time consuming file system checks.And also uses battery-backed up non-volatile RAM (NVRAM) to avoid losing any data input/output requests that might have occurred after the most recent consistency point.

RAID: Redundant Array of Inexpensive Disks

The storage appliance uses either RAID Level 4, which stores all parity information on a single disk, or RAID-DP, which stores all parity information on two disks to protect against disk failures.With RAID 4,when a disk fails the parity disk is used to recalculate its contents automatically after the disk is replaced.
The parity block is basically the sum of the blocks in the stripe.
Parity=15 Data 1=4 Data 2=5 Data 3=6
The RAID system uses EXCLUSIVE-OR instead of addition and subtraction.RAID-DP can recover from double disk-failure while allowing the RAID group to continue serving data and recreate data lost from up to two failed disks.

Other RAID levels

RAID 3,uses a single parity disk and systems often keep the disk heads synchronized to reduce latency.So RAID 3 is a good fit for applications that require a very high data rate for a single large file(super-computing and graphics processing).
RAID5 is like RAID4, but it cycles parity among all of the disks in the array.The primary disadvantage is that it is not practical to add a single disk to a RAID5 array because to add new disks,a new array must be added.