Visualization, whether it is artistic, conceptual, informational, or scientific, involves creating and moving around a lot of data. We have covered Fusion-io in the past simply because it looks like great technology that can be used in that process. Fusion-io makes solid-state drives (SSD) that fits into a x16 PCI Express 2.0 slot. The beauty of using a x16 PCI Express 2.0 slot is that it enables Fusion-io to obtain a higher bandwidth.
Fusion-io has recently announced the availability of its new ioMemory module. By using the latest in MLC NAND Flash, they are able to double the capacity of its products. This gives them the ability to have up to 1.28 terabytes (TB) of capacity. But capacity is not everything. This latest memory also allows them to have 285,000 sustained IOPS with under 25 microseconds latency.
Single point of failure? You mean like a RAID controller? Fusion’s card takes the slot of a RAID controller and has 4x the MTBF of a RAID controller. Oh and at the end of the day, everything communicates with the CPU via the system bus, which is generally PCI Express.
@Bob you can always use SAS/SATA SSDs in redundant mode. There are MANY alternatives if single-point of failure worries you in Fusion-IO.
@ Tom C
All systems have single points of failure. Single servers are seldom used in mission- or business-critical applications, specifically to protect against both hardware and software failures that cause down time. Storage devices, including solid-state storage are generally the least of your worries in this area.
The same rationale applies to the comment regarding hot swappability. The vast majority of components in servers are not hot swappable. This is either because their failure rates are insignificant, or because redundancy (hence swappability) is done at a higher-level (e.g.: the server-level) in the solution architecture.
I note that the failure rates for Solid-state Drives are quite low. MTBFs are often in the millions of hours. Adoption in the Enterprise is very strong both for performance and reliability reasons.
Unfortunately with both Fusion-io and upcoming Seagate PCIe card you still have a single point of failure. If a PCI card fails, even with two PCI-e cards as RAID 1, you will end up with kernel panic or blue screen! Bottom line, sacrificing availability and reliability for simple performance is not an Enterprise solution. Of course you can have multiple servers with multiple cards, but you just doubled your cost! Great for my home PC if it was bootable and was cheaper than my car but not ideal to risk my datacenter! If it is not hot-swappable is not for Enterprise!
Oh yeah, try running these cards for a week under any real write benchmark! Spikes in writes will make you think twice
@ Chad
The card in the picture is the ioDrive Duo and that card supports X8 PCI Express, which is wide enough to achieve 285,000 IOPS at very small packet (or block) sizes.
I think they use x4 slots. They aren’t fast enough to need x16, and those are harder to get in quantity on most servers anyway.