Whither HPC, HPE?

HPE is acquiring Cray Inc. Almost 3 years ago, HPE acquired SGI. Back in 2017, HPE partnered WekaIO, and invested big in the latest Series C funding of WekaIO just weeks ago.

Cray, SGI and WekaIO are all strong HPC technology companies. Given the strong uptick in the HPC market, especially commercial HPC, we cannot deny HPE’s ambition to become the top SuperComputing and HPC vendor in the industry.

HPC Stepping over each other now

As an industry observer, and an outsider, I cannot get over the fact that there are already overlaps with all 3 companies. The commercial HPC market is now driven with new workloads of deep analytics, machine learning and AI. The traditional engineering/scientific market is still strong but the attractive lure of HPC, is without a doubt, this new commercial HPC workload.

Cray has been strong in the traditional HPC market but have, in recent years, moved into the commercial HPC space. SGI, which the SGI 8600 competes with the Cray Shasta platform in the traditional HPC market, is also in the commercial HPC space. WekaIO, the young start-up and the world fastest parallel file system, is also in the commercial HPC space.

Where do we fit all 3 of them – SGI, Cray and WekaIO? The overlaps in the tight commercial HPC markets are quite glaring.

What about others?

Mergers and acquisitions are part of parcel of the technology market place.

HPE’s HPC competitor, Data Direct Networks (DDN), acquired Tintri in a fire sale last year and announced its intent to acquire Nexenta 2 weeks ago. The DDN agenda is more clearly demarcated with Tintri and Nexenta, because both companies are different in their offerings to DDN’s technology portfolio. Tintri is in the VM storage space, while Nexenta has fashioned itself into the telco space in recent years.

But even with this factor, customers still want to know where DDN is going with its acquisitions. The roadmap for Tintri is still vague, according to a Techtarget article last week. What can we expect from HPE?

HPE, are you really focused?

HPE already has a strange and confusing technology strategy. After “divesting and investing” its Enterprise Services Division 2 years ago to (or is it with) Computer Sciences CSC and now known as DXC Technology, this was followed by an “alliance” with Microfocus months after that. But from the outside, it appeared that the whole HPE Software was shipped off to Microfocus, dusting its hands clean.

Then there is the HPE Complete strategy, an “end-to-end infrastructure 3rd party solution” portfolio. In my view, from a storage perspective, they have

• Datera – scale-out primary block storage
• Commvault Software – what happened to Data Protector?
• Zerto
• Cohesity – aren’t there similarities with Commvault Hyperscale?
• Scality – HPE invested USD$10 million in Scality in 2015 but memories of HPE’s own Helion Cloud lingers
• Qumulo – scale-out file storage, OEM deal
• Hedvig – scale-out, multi-protocol
• Cloudian (reseller agreement in EMEA only) – object storage (competes with Scality) with NAS (Qumulo competitor)
• WekaIO – scale-out parallel file system

I have highlighted the word scale-out from the 9 companies above, and it appeared 4 out 9. If this entire HPE solutions world is confusing to me, imagine how an HPE solution architect will have to go through to put this forward to a customer.

And then there is that Nutanix thing, a recent global partnership to expand HPE’s Greenlake offering. I raised a bit of a hoo-hah since I am a big fan of HPE Simplivity. In my books, HPE Simplivity is the best hyperconverged platform in the market today. How would this Nutanix move reverberate within the HPE Simplivity business unit is anybody’s guess.

Since voicing out my confusion a few months ago, I have received a share of naysayers from the HPE community. “Customer choice” was the loudest I have heard. I would agree it would be the customer’s choice, if the customer has clarity of HPE’s strategy in its offerings. Not when the solution landscape looks like a patched up quilt work of sorts.

A clear HPC roadmap and strategy required.

The HPC market is a mature market and customers are in it for the long term. What they will invest in will be something that will bring them continuity and value for their business and operations (The operative word here is Continuity). And I seriously want to see HPE getting its HPC portfolio act together. A clear roadmap must be communicated as soon as possible to avoid both internal and external fall out.

With the usual rhetoric of “customer first” and “customer choice”, HPE must also address its culture of innovation and keep developing forward. The basket of 3rd party solutions, partnerships, OEMs, resellerships and “spin-in and spin-out” is probably good for short-term and maybe mid-term plans, but HPE should be more than that.

After all, we don’t a case like HP Ibrix (clustered parallel file system) getting shoved into the closet and never to be heard again.