HPE had its own little conference called HP Discover this year, in the Excel Center in London. First a little disclaimer. I was invited to attend by Calvin Zito @HPEStorageguy (so thank you for that). That said, it will not affect my independence regarding technology. Now a little aside: one of the oddest things for those of us who are UK based is that even though this was a UK-based conference, we had to use European-to-UK power adapters.
The Good, the Bad, and the Ugly
This is a surprisingly well-organized conference. That said, it should be, given the price they charge for tickets. It is not a VMworld; it is not deep technical. This conference is aimed front and center at decision makers, not influencers, as can be seen by the number of suits in attendance. There are no pizza-stained AC/DC T-shirts here. The briefings veer more to strategic than technical. This is not a bad thing, as the material is targeted to the intended audience.
Of note, there did not appear to be a massive solution exchange, and the stands were spaced well apart. It did not look empty: the spacing was more elegant. When you were talking to the personnel at the stands, you felt it was a more intimate conversation. There was not the feeling of being jostled from A to B, or that you were in a meat market. This is where the conversations got interesting for me, because the average stand-to-attendee interaction was at a higher, more strategic level. The techies doing the demos and displays were interested and, more importantly, available to go deeper.
What Did I Learn?
The main thing I learned is that HPE hasn’t got a clue what it is anymore. Its messages are fractured and discordant. Now, this does not mean that its management team is floundering. The briefing we received from the CTO was cogent, and I got the feeling that in two to five years, things will be interesting there. But the front line is confused. Personally, I attribute this to marketing and sales still trying to figure out their space in the world after the splits and spin-merges. I would like to return to this next year once the ripples of the split bomb have flattened out.
This is not to say that it is stagnant. There were plenty of interesting announcements. “The Machine,” a research project out of its labs in Fort Collins that was intended to be the next iteration in computing engines, was in a bittersweet moment. It was premiered as a prototype in London, but it seems that HPE has shelved intentions to commercialize the product. Nevertheless, it has some interesting innovations, and these are filtering down into HPE’s other commercial products.
We are already seeing Persistent RAM (first-generation NVDIMMs: standard DIMM technology backed by NAND flash) in ProLiant Servers. This is interesting tech, and it is available now. I will be writing more about it in a later article.
HPE announced a further divestiture. It has hived off its OpenStack and Cloud Foundry contributions to SUSE. (Interesting fact: SUSE is owned by Novell, which had its assets acquired by Micro Focus, to which HP sold Autonomy.) It will, however, be retaining its seats on the foundation boards.
This means that HPE has effectively split into two separate companies: HPI and HPE. Large swaths of HPE, in terms of software and technical services, have gone to Micro Focus and CSC. However, what is interesting is that Micro Focus and CSC were not outright sales but rather mergers; HPE shareholders hold 50.1% of shares in each of those two companies. It is a small distinction, but I must say shareholder; that said, I am sure that there are HPE people sitting on the boards of those two companies to offer direction. These spin-outs/mergers/sales are quite complicated and will be the subject of their own article, but they have delivered a large amount of potential shareholder value.
HPE has rebranded its network division as HPE Aruba. This makes sense, but its product set still does not. Its network offerings are disjointed and consist of products developed along three separate tracks, resulting in three completely different interfaces to manage them. This needs to be sorted. There is a reason why Cisco has dominated the networking arena for so long. I am still unsure about its SDN/NFV story.
Summary
The ProLiant, still one of the most popular x86 server brands in the world, is still being developed; whether this is important in the cloud-led COTS world we now appear to be living in remains to be seen. Hybrid cloud will help, but with more intelligence being held in software, and a throwaway ethos entering the mindset regarding servers, containers, and services in general (think Pets vs. Cattle), HPE has a battle over hearts and minds to win. That said, HPE does have its eye on the next generation of compute. “The Machine” is evidence of this. The technology’s potential for managing very large datasets is impressive, and even though it is not going to be productized, the research is being trickled down into products. Moonshot is evolving: we were demoed new composable hardware that is being aimed at the IoT market. HPE does have answers, but it is struggling to find the questions.
HPE is at a pivot point. Like a lot of other mainstream companies, it is fighting to remain relevant in an industry dominated by a desire to get wet with cloud, and fog, and then to forget about servers completely with the new paradigm of serverless. That HP does have excellent servers in its ProLiant and nonstop products, good storage answers in 3PAR and StoreVirtual, and passable network answers, is without question. However, from this conference, it appears that this is not all. HPE seems to be floundering; its strategy is currently unclear. I do believe it will survive this storm, but it is very uncertain what form it will take. Perhaps by next year’s US Discover in Vegas, we will see a clear and coherent strategy emerge from the waves and breakers that are rippling through the company due to the split.