EMC spends more cash on Flash

I do not believe much needs to be said on this acquisition. We all know who EMC is and what it is famous for. ScaleIO last year (2012) came out of the stealth mode and recently got acquired by EMC for something like $200m to $300m depending on the source. Looks like portion of the team came from XtremIO that was couple of years acquired by EMC (see the pattern?).

Continue reading

Fusion IO beefs up with another storage acquisition

Fusion IO finally announced acquisition of NexGen storage after a few month NexGen was quiet comparing to its SMB targeting storage competitors like Nimble. There were signs and rumors that were putting FIO as one of the most promising candidates for the acquisition, and it happened.

With this new addition to Fusion IO portfolio it has a clear path into Continue reading

FusionIO acquires ID7

Today FusionIO acquired UK based ID7 provider for Software Defined Storage. ID7 is a driving force behind the Generic SCSI Target Subsystem for Linux. It enables the following functionality: replication, thin provisioning, deduplication, high availability, and automatic backup on any Linux server or appliance (http://scst.sourceforge.net/index.html).

ID7 worked with FIO on ION which now provides more context for the last conversation that we had with FIO on SDK where Stephanie indicated that ION implementation has switched (no more details were given). ID7 joined the ION Acceleration effort where they will continue driving this effort (http://www.prnewswire.com/news-releases/fusion-io-acquires-id7-198762381.html).

It is expected that the products developed through this merger will target SAN systems users from mainstream storage suppliers, most of which are high-end. However, “we won’t aim at the highest end users,” says Flynn. ”Facebook, Apple etc use the application database to manage things. But hosting and cloud providers, for example, are often forced to buy proprietary systems. Software-defined systems can offer them high-end systems without the high-end price.”

With the ID7 acquisition, Fusion-io will pick up both general and specific software expertise. While ID7 develops storage software for iSCSI, Fibre Channel and Infiniband, SCST was originally designed for both SCSI and Linux. That represents one of the smaller parts of what IDC calls the Open SAN market, which consists of FC SAN, FCoE, iSCSI, and others. (About $3.2 billion of iSCSI storage arrays were sold in 2012, IDC said, according to private figures cited by EMC marketing CTO Chuck Hollis.)

Since SCST’s initial development, Fibre Channel and iSCSI support have been added, as well as support for Fibre Channel over Ethernet. And SCST supplies quite a bit: replication, thin provisioning, de-duplication, high availability, and automatic backup on any Linux server or appliance. (For more on building an SSD disk array over Fibre Channel using SCST, see Marc A. Smith’s blog post on the subject.)

Fusion-io said it would continue to maintain the open-source version of SCST, as well as contribute to the open-source distribution as it continues development. The ID7 team, including the primary SCST developers, will join Fusion-io but work from their current locations (http://slashdot.org/topic/datacenter/fusion-io-buys-id7/).

“Across Linux and Oracle subsystems, Fusion-io has embraced the need to support open source contributions to innovation in the storage stack,” said Vlad Bolkhovitin, Founder and Primary Developer of the SCST SCSI Target Subsystem for Linux. “By joining the Fusion-io team, we are pleased to be able to provide SCST with more engineering resources and talent to support the development of the open source and commercial SCST subsystems as we meet the needs of Fusion-io customers around the globe.”

In conclusion this acquisition positions FIO not only as a provider of Flash for homegrown scale out solutions engineered inside the Facebook and Apple datacenter but also play a key role in the datacenter providing high end storage with ION directly against EMC’s XtremIO.

Flash Wars. EMC comes out with Xtrem guns

As expected sooner or later that had to start since EMC’s 400M acquisition of Israel XtremIO, and now it begins.

Pulling of VFCache from USD (Unified Storage Division) and creating Flash Business Unit that now includes VFCache renamed to XtremSF (LSI “although EMC is not revealing who they are” ;) ) and XtremSW (home grown VFCache software) along with XtremIO.

I was always questioning where middle-ware product similar to iON would actually make sense, and sure enough it does not seem it does: “Thunder server networked flash cache product built from VFCache cards has been discontinued”.

EMC is showing very interesting results “Fusion-io being able to run at around 115,000 IOPS with a 70/30 read/right mix of 4K blocks and a queue depth of 256. The XtremSF 2200 delivers around 215,000 IOPS in the same situation.”

“EMC has also tested the XtremSF and Fusion-io cards over a greater than 5 hour run period and found its cards quickly outperformed Fusion-io cards with 4KB random write IOPS and eventually overtook the Fusion-io card with 64KB sequential writes in MB/sec terms.”

They will be leading on the VMWare integration as well.


Of course the response from FIO is expected. Inheriting EMC’s possession in fast cars FIO creates a video of FIO vs E brand (not hard to figure our what goes after E):

Who is who in “Flash Wars”?

  • EMC – a potential powerhouse but it must link its software to backend arrays as soon as possible or server makers could freeze it out.
  • Fusion-io – software is the key and Fusion-io already knows this.
  • IBM-TMS – we’re still waiting for IBM to add TMS flash to its servers and get busy on software driver front.
  • Intel-Micron – it owns foundries, has distribution channels and it’s getting software.
  • LSI – no foundry link-up but potentially has a good distribution channel although it needs software.
  • OCZ – no foundry link-up and in crisis.
  • Samsung – can sell PCIe flash on the back of memory and flash chips to system manufacturers so it’s potentially strong – but it needs software.
  • Seagate-Virident – needs a flash foundry tie-up and that could be Samsung.
  • STEC – its high-flying SSD business stalled, although it’s now recovering, but it has no foundry tie-up so question marks remain.
  • Toshiba-SanDisk-Violin – a potential powerhouse.