HP announced some weeks ago, during its annual HP Discover event in Las Vegas, the last 3Par StoreServ model, named 7450, and completely based on flash disks.
This article is not willing to be another product review, I’m not a storage expert by any mean, at most I can be considered a virtualization expert who uses several storage systems. You can find many reviews and analysis around the web, and I liked for example this one from Chris Evans and this from Enrico Signoretti.
So, as a Storage “user”, I would like to write some thoughts I had about this new launch from HP and how it could influence the market, and how in my opinion now HP has a new “weapon” to fight the battle with many startups who came out in the last years in the all-flash storage market.
Make or Buy?
First of all, a plaudit goes to HP for balieving and investing in the 3Par technology.
Usually, when a big corporation is late on some technology if compared to some startups, the quickest and easiest move is to use part of their huge funds to acquire the startup itself. With this move it gains two advantages at the same time: it takes away a competitor, and it gains all the technology the startup had developed. We have seen this behaviour many times, in the same market segment for example EMC lately acquired ExtremeIO.
HP instead chose a different approach. In the endless debate about “make or buy”, HP bet on make. Last dicember I was invited as a blogger to the european edition of HP Discover, in Frankfurt. During a roundtable with David Scott, VP di HP Storage, a question came out: “are you going to acquire external technologies in the flash array market?”. And David promptly replied (those are not the exact words…) “No, we think HP has all the needed technologies to develop this kind of products in house. You’ll see”.
Well, the result is the new 3Par 7450.
Big Vendor vs Startup
With this announcement, now HP has in its portfolio an all-flash array exactly like many startups. Not a simple announcement about a product that is not yet available, but a product ready to be sold and shipped. In previous months, many startups on their corporate blogs had many articles criticizing and having fun of announcements from EMC (ExtremeIO, ViPR) or NetApp (EF540, FlashRay). I do not like this kind of articles, they appear to me as rude defensive moves with no added value to those who write them, but nonetheless they were right in criticizing the big vendors in announcing products with so much advance, products there are not really existent at the time of the announcement; while them (the startups) already had available and mature products. Now, the HP announcement has received no comment from any startup. A good sign that this new HP product had hit hard.
On of the ideas that came out, is that there is an objective risk in acquiring technology from a startup, so to make it “valuable” to the end user, there should be a remarkable technology gain when compared to the same (or similar) technology of the big vendor. No, if we try to apply this idea to the topic “3Par 7450” vs “all-Flash Startups”, I see the comparison between them moving in favor of HP.
First of all, the underlying technology. To me, a all-flash array is needed mainly for performances. And with 550.000 IOPS and under millisecond latency, the new 3Par performances are comparable with those of some startups (Pure Storage for example says they can offer 400k IOPS with same latency). For sure, doubts about the way those IOPS have been calculated can arise, but nonetheless we are talking about comparable storage systems.
Price per GB: 3Par has always been known for its ASIC, a dedicated chipset designed to data reduction. When applied to SSD disks, the ASIC allows (just like the technologies of many startups) savings by reducing the number of writes on the SSD, and so lowering the price per GB of the array.
In a way, now probably HP has even an advantage position if compared with the startups. The most relevant issue is about the “enterprise” features the storage has. Startups often lack some enterprise feature the large companies require, the biggest of them beeing replication between storage arrays (and by consequence, the support for example of VMware Site Recovery Manager). All the startups has those feature in their roadmap (but none of them would say no we do not…), but really few of them have those features NOW. On the other hand, the new 3Par is a 3Par, so it has from day 1 the complete feature set: snapshots, replicas, peer motion, federation…
To sum it up, HP has really reduced the gap between 3Par and the all-flash startups, or maybe it is even gain advantage on them. And so, if a startup has no more a remarkable technology advantage when compared to a big vendor, is it still worth the risk? Or maybe the enterprise customer is going to trust the big vendor more than the startup?
I’m waiting for the months to come, to see how this market is going to evolve.