Phone: (03) 8679 0319
 
Email: info@iomax.com.au

Performance.

The performance that IOMAX can deliver to your servers and applications is both class-leading and easily scaled. This means we can provide a quick solution for a sluggish database server right up to enabling organisation wide access to information in real-time by super-charging SAP Business Intelligence servers. How?

The key is IOPS or input/output operations per second. The number of IOPS available is crucial to the performance of applications and servers hosting databases, mail-servers, analytics packages and other enterprise applications.

IOPS are input/output operations per second and is a function of the bandwith and the latency of the storage device. Think of a road:

>> A wider road with more lanes is better (greater bandwitdh).
>> A road which has no traffic lights is better (lower latency).
>> A road with both becomes a freeway and is the best!

IOMAX products are designed from the ground up to deliver very high IOPS - each ioDrive has bandwidth of at least 750mb/s with latency of less than 50 micro seconds.

The following chart shows how IOMAX products measure up:

As you can see, IOMAX products really do provide extremely high levels of performance and scale almost linearly.

SPC-1 Test Results.

We compared an IOMAX ioDrive powered solution to other enterprise-class solutions on the SPC-1 benchmark. Here are the results:

Reliability.

NAND use to have a bad name - but no longer.

IOMAX prodcts utilise a comprehensive approach to data integrity that have solved the reliability problems of NAND. We utilise class-leading data protection and reliability measures that make our solutions even more reliable and long lasting than ageing disk technologies.

How have we done this? In three ways:

So in summary, we use RAID 5 on board, have a decidated parity chip, placed 20% extra NAND on-board, apply an advance wear leveling alogorithm and have the highest levels of error correction.

The combined Result is a MTBF of 25+ year,an expected lifetime of over 10+ years (based on 1TB per day of read-erase cycles) and the ability to gracefully recover from failures at all levels.

>>Click here to see download the full paper on reliability...

Case Study.

Wine.com’s SQL Server system was running at capacity and plagued with poor performance and high latency. They estimated the poor performance of their system resulted in an estimated loss of 15% of sales. But, by using the FusionIO ioDrive, they achieved large performance gains for less money and saved on other costs as well including power and server hosting.

The following is a summary of the results that were achieved:

>>Click here to download the full Wine.Com case study...

Technical FAQs.

Doesn't NAND flash have a write limit? How does that effect the lifetime of the ioDrive™?
NAND flash has a limit on the number of writes that can be done to an individual cell. The particular limit depends on the type of flash used. For Single Level Cell (SLC) NAND, the limit exceeds 100,000 writes to a cell, whereas for Multi Level Cell (MLC) NAND, it is on the order of 10,000 writes. Hence, in order to exceed the limit of a single 80GB ioDrive™, you would have to write almost 80PB (Petabytes) of data. Streaming data at 800MB/s to the card, it would take you 3.4 years of writing data non-stop to exceed the SLC limit.

How does the data reliability of the ioDrive™ compare to a traditional hard drive? The ioDrive is far more reliable than a hard drive as it relates to data reliability. In both hard drives and ioDrives, Error Correcting Codes (ECC) are used to ensure that any minor failure doesn't cause data loss. Fusion-io uses strong ECC algorithms to ensure the safety of your data.

Is the ioDrive prone to catastrophic failures similar to a hard drive?
Most catastrophic failures in hard drives are the result of mechanical failures, causing total data loss. Because the ioDrive has no moving parts, it isn’t subject to the same mechanical failures of traditional hard drives. When a NAND chip on an ioDrive™ begins to fail, the data is simply written to an adjacent NAND chip. The result is zero data loss with only slight reduction in overall ioDrive storage capacity. Furthermore, the ioDrive’s RAID capabilities protects from controller or other major (but extremely rare) types of failures.

Is it possible to RAID multiple ioDrives™?
To the host operating system an ioDrive appears as a normal hard drive, so users are able to leverage any software RAID solution that uses the host Operating System’s native drivers. The Operating Systems volume manager performs the RAID function.

Can you truly replace RAID disk applications requiring high rate IOPS and life time in the 250,000 hr range?
An ioDrive has a product life in excess of 250,000 hr - even when performing continuous writing to the NAND chips. Moreover, users can expect to achieve performance in excess of 100,000 IOPS by aggregating multiple ioDrives in a RAID configuration. Hence, an ioDrive-based system is quite superior to a RAID disk array.

What is the latency of the ioDrive?
Small packet reads take approximately 50us. On writes the Operating System level write buffers are freed in just a few tens of microseconds once the data has been transferred into the NAND registers (buffer). It takes approximately 200-300 microseconds for the data to be imprinted in the NAND chip by a flush/sync directive or once the page is filled with write data.

Does the ioDrive require a back-up power supply to protect from power outages?
The ioDrive uses non-volatile NAND flash as its storage medium. Hence, when data is stored in the NAND chip, it is guaranteed persistent.

Will the ioDrive be a bootable device?
This feature is in the development pipeline but not yet a reality

Can the ioDrive be partitioned like that of a hard drive?
Yes, users can partition the ioDrive. This is done in the same way as any other block device; by writing a partition table into the raw block device (using fdisk). Then users can use the partitions how they please - including as storage cache.

Is the ioDrive affected by fragmentation like a hard drive?
Users don’t need to worry about fragmentation with the ioDrive. A file can be scattered randomly across an ioDrive and it will still achieve full performance – this benefit is due to the ioDrive not having to seek a mechanical read head.

What methods are available for accessing the ioDrive?
The only method for accessing the ioDrive is by block device. The ioDrive looks just like a hard drive to your OS. You can access the drive directly, format it with a file system, or even set it up as a swap space.

Does ioDrive work with my motherboard?
As Fusion-io cannot test every motherboard the following guidelines should be used. Any quality motherboard that has a PCI-Express x4 or wider slot should work at full performance. Any 4x physical slots with only one lane will work if slot is properly built to spec.

Does ioDrive work with different PCI-E slot widths?
The ioDrive currently uses x4 signaling and x4 physical connector. It can be plugged into wider slots such as x8 and x16. The ioDrive will also work in x4 physical x1 signaling slots but with degraded performance. We cannot guarantee that using a x4 (or larger) physical to x1 physical adapter will work unless they are PCI-E SIG compliant.

Does the ioDrive require a RAID controller?
The ioDrive does NOT need a RAID controller as it is directly connected to the system PCIe bus. Multiple ioDrives can be configured using the host Operating System’s logical volume management to do RAID 0,1,5 etc between ioDrives. Such an implementation is MUCH faster than using hardware-level RAID controllers.

How does an ioDrive’s RAID functionality compare to a traditional disk-based RAID solution?
The ioDrive significantly outperforms the highest performance hardware-based RAID controller on the market.