Reason #1 ZFS is the Best – Copy on Write File System
The copy on write architecture is the foundation of what makes ZFS so awesome. With a copy on write architecture no data is overwritten which means a write no matter how large can be written as a single atomic transaction. An interrupted write is never committed because the pointer to the new uberblock, which is the last action of a write is never written. This same copy on write architecture enables several other features such as snapshots and clones to happen with minimal resources. By leaving a pointer to the top of the tree for the old data ZFS enables a snaphot, a pointer to a valid copy of data at a particular time. A clone is simply a duplicated pointer which can now move independently from the old pointer.
Reason #2 ZFS is the Best – Self Healing Architecture
ZFS use a top-down, end-to-end 256-bit checksumming system (SHA256 and Fletcher4). Unlike other checksummed storage systems, ZFS stores checksums not within blocks, but with the pointers to blocks, all the way up the system metadata tree. When a given block is read, its checksum is calculated and compared to the checksum stored in its parent block pointer. By comparing checksums throughout the metadata tree, ZFS create a “Data Integrity Field” that detects both noisy and silent errors, including:
- Bit Rot
- Phantom Writes (undetectable by hardware RAID)
- Misdirected Reads and Writes (undetectable by hardware RAID)
- DMA Parity Errors (undetectable by hardware RAID)
- Driver Bugs (undetectable by hardware RAID)
- Accidental Overwrites (undetectable by hardware RAID)
In day to day use, when a damaged block is read and detected, ZFS automatically repairs itself by copying data from a known good copy. Self-repair occurs on the fly. ZFS also runs a preemptive “scrubbing” process, similar to ECC memory scrubbing, that operates on live and in-use file systems at low I/O priority with minimal performance impact.
Reason #3 ZFS is the Best – Hybrid Storage Pools
ZFS’s hybrid storage pools enables users to mix drive technologies to achieve unprecedented performance with a cost that is lower than any traditional storage solution. Using RAM and SSD’s in front of traditional spinning disks ZFS can deliver 100’s of thousands of IOPS using a few 7200 RPM SAS disks. A traditional storage solution requires 1,000’s of 15K SAS drives. Even if you could afford all those drives do you have the power, space, and cooling to host them? ZFS is great for streaming data to with intelligent pre-fetch caching.
What is the Best ZFS Appliance for Your Storage Needs?
BrickStor is the leading ZFS storage appliance. BrickStor scales from 15TB’s up to several PB’s.BrickStor can be configured to meet the needs of a wide range of customer sets even from within the same storage appliance. BrickStor is a total turn-key ZFS solution. Unlike our competitors, BrickStor is developed, tested, and delivered to you the customer as an end-to-end Enterprise Class Storage solution. You should want a ZFS storage solution; you should want BrickStor.
