3PAR: Growing Beyond a Niche Company?

On a call yesterday, 3PAR CEO David Scott and VP of Marketing Craig Nunes met with industry analysts to review quarterly results. Although I typically enjoy earnings calls, this call was much more interesting than Tuesday’s call with Wall Street analysts who seemed intent on dissecting 3PAR’s change in revenue recognition policy.

Scott shared with analysts that 3PAR’s growth is running at an 11% year over year and 7% sequentially but he provided some other interesting perspective on these figures. The most telling for me was that over a 2-year period 3PAR is growing at 52% while the high end revenues of players such as EMC, Hitachi and IBM are declining in the low single digits.

I asked Scott about the comparison to the high end (e.g. Symmetrix, USPV and DS8K) and he said that 3PAR predominantly competes with these high end products and not the midrange systems typified by EMC’s CX line. I was somewhat surprised by this but 3PAR has been saying for a while that it competes directly for so-called tier-1 storage opportunities. I’d like to dig into that a bit more an try to understand if 3PAR is predominantly taking out, for example, much older (e.g. N-2) Symmetrix systems or if it is truly vying for upgrades to new VMAX opportunities.

I know what the answer is from 3PAR but I’ll be paying more attention to this. It’s clear 3PAR is a serious player in the high end (however you want to define that) and is moving up the value chain, competing to support more mission critical applications. I just hadn’t realized to what degree.

It’s also clear that at these growth rates, over the 24-month period, 3PAR is gaining share. Scott says to continue that pace it will have to invest in feet on the street and channel, particularly in non-US sectors which are growing considerably faster for 3PAR. While it is growing from a small base, at roughly $200M in annual revenue, 3PAR has a lot of upside. You can correctly point out that EMC throws off more free cash flow in one month than 3PAR generates annually in revenue (I know…EMC is a bloody cash machine) but the company is starting to make inroads into the high end and will likely continue to grow faster than the market.

The Wall Street Journal keeps reporting that 3PAR is a buyout target but I sure hope 3PAR’s board let’s the company grow more before it sells itself.

Invoking the F-Word and Other Updates

Scott talked about storage federation, presumably in an attempt to provide some context leading up to EMC’s expected federated storage push at EMC World following up from Pat Gelsigner’s March teaser. Scott stressed three things about storage federation:

1. It’s different from virtualization. It allows separate peer, self-governing systems to act as a global whole (versus a hierarchical approach).
2. Automomic storage will be a critical enabler to federation as stuff gets moved around the virtualized network.
3. Storage federation is following a natural evolution from data center–>metro–>globally shared storage.

Craig Nunes then talked about 3PAR’s Adaptive Optimization, tiering and SSD uptake saying that around 2% of 3PAR’s total capacity shipped was SSD’s and the early interest in Adaptive Optimization is strong– no surprise there the stuff is cool. Nunes also talked about Oracle ASM integration with 3PAR’s thin provisioning and how 3PAR does space reclamation in-line using cache and won’t incur a post-process penalty for what I like to refer to as garbage collection (I’d never make it in marketing). Nunes threw out the stat that for every TB of 3PAR supporting Oracle, there’s the potential to unlcok 550GB of capacity.

3PAR also talked about new VMware vCenter plug-ins that are simple and allow point-and-click backup of VM’s and VMDK’s to InServ but it didn’t give any details on its VAAI integration (vStorage APIs for Array Integration). I always ask storage suppliers where they are with VAAI because that’s the down and dirty stuff that will really solve the storage problems associated with VMware. VMware breaks storage and VAAI integration is the fix. Here’s a piece Floyer wrote about this that explains the storage roadmap for VMware which starts with VAAI integration. 3PAR is working on this like every other storage vendor but the deck is stacked for EMC (because it owns VMware) and a little bit for NetApp (because VMware knows NetApp is big and wants a second big storage partner). Lots of vendors complain that VMware is hard to work with and many say they don’t even have the most up-to-date SDK for VAAI.

Nunes threw out another really interesting stat that evidently comes directly from VMware. It says that NAS infrastructure requires five times more CPU costs than SAN. I haven’t found the documentation yet but I did find this statement from VMware in a performance best practice paper:

Fibre Channel SANs typically yield higher performance than NAS and iSCSI. Fibre Channel SANs do not suffer from the congestion and oversubscription problems as readily as NAS and iSCSI because of the Fibre Channel protocol.

I’ll get the actual reference from Craig and update this post with the link. The implication is that on a per TB basis, NAS will support far fewer virtual machines per ESX.

The other fun fact provided was a supporting statistic from a third party research firm (I presume IDC or ESG ) that 81% of VMware storage is SAN-based storage. That is an astounding figure when you think about it. Again I’d like to better understand what’s behind those figures.

Bottom Line

3PAR: Growing, more than a tier 1 wannabe, focused, independent (hopefully for a while), not profitable enough and still too small to move markets by itself. But buyers are starting to take notice and the company has for several years now been considered the gold-standard (along with Compellent) for simplicity and ease-of-use – two concepts with which the storage industry hasn’t been historically associated.