Sysadmin's 2015 review part 2 With 2015 drawing to a close and 2016 about to begin it's time to reflect on the fact that the world never stops changing. The tech industry certainly changes constantly, and so here's one sysadmin's view of the industry's movers and shakers.
In part one, I took a look at Amazon, Oracle and Microsoft. Here in part two I am going to attempt to decode the bizarre and curiously intertwined dance between Cisco, HPE and Dell.
In case you missed it, Dell announced it was buying EMC. Having recently done a teardown on the acquisition, I'll skip discussing the full implications here. The acquisition of EMC's component parts – EMC II, VMware, RSA, Pivotal and Virtustream – offer Dell plenty of ways to diversify its revenue streams and don't bode particularly well for the competition.
The key here is revenue diversification. Dell went private because its previous approach to things wasn't working. It is spending an unbelievable amount of money buying EMC ($67bn) so that it might have its fingers in a whole lot of different pies, meeting a number of needs spanning the entirety of its customer base's requirements.
More critically, Dell acquired research and development capacity. EMC's various sub components are very engineering-heavy organisations. Some of the best and brightest work there, and they work every single day to advance the state of the art in IT.
When Dell went private it did so because Michael Dell felt that the short-termism of Wall Street was crippling its ability to succeed. Private-equity firm Silver Lake clearly agreed because it bet its future on Dell (the man)'s vision.
Microsoft also agreed, because Microsoft loaned Dell a pile of cash to help the deal go down. None of this should be surprising; Microsoft also has massive investments in research and development and it is this technological leadership that has kept Microsoft relevant in the face of increased competition from all sides.
Michael Dell is essentially executing on Dave Packard and William Hewlett's legacy: research and develop complex systems in house, and then provide whatever technology it is that people seem to need. This is definitely a long-term play and it remains to be seen if Michael Dell's patience really is more robust than that of Wall Street.
If Dell is busy being what Hewlett-Packard Enterprise (HPE) - the enterprise wing of HP spun out in 2015 - was, then HPE is busy trying to succeed where Dell failed. Having jettisoned the endpoint and printer side of its organisation, HPE is tightening its focus on enterprise products and services.
With the exception of Openstack, the old HP had spent more than a decade killing off its research and development capabilities. Every time HP acquired a company that threatened to provide HP with new and innovative R&D, HP found a way to drive the engineers right back out again.
HP famously curtailed R&D and the waves of job cuts will surely have been seen as an exodus of talent along with the trimming of the excess staff.
So HPE isn't going to be dependant on R&D. In this writer's view, it's going to become a QA specialist. HPE's game plan will be to take someone else's software, put it on hardware made by a different someone else, test it to make sure it doesn't do anything strange, and then sell it at a premium.
HPE's recent reinforced Azure partnership with Microsoft is an important example. The newly announced hyperconverged cloud in a can Azure box is only the first step. As Dell builds a diverse and integrated empire, HPE is looking for love outside of Dell subsidiaries, such as VMware.
In essence, Dell is bringing in-house all the pieces needed to build infrastructure endgame machines, while HPE is going to assemble them from bits that other vendors invent.
The exception to this is, again, Openstack. HPE's efforts with its Helion private cloud have led to some interesting customer wins.
HPE is killing its public cloud. But it is pushing forward with private cloud deployments. Meanwhile, it is ready to start competing against whitebox vendors on their own turf. A playing field, it's worth noting, that is completely bereft of margin.
Dell couldn't make money fast enough to please Wall Street by repackaging other people's stuff and selling services on top. Can HPE? Not having the profitless endpoints to worry about will certainly help, and there does seem to be money to made with this "data centre in a can" idea.
I suspect that everything will hinge on the success or failure of HPE's Helion Cloud. With the margin dropping out of tin shifting, storage, networking and even services, all that's left is to "go big or go home".
Cisco is a company in transition. Cisco relies on customers buying its networking gear to survive. In fact, not only must customers buy its gear, they need to buy it in bulk ... and above all they need to buy Cisco exclusively.
Unfortunately for Cisco, the days of its networking predominance are past. The industry as a whole has become wise to the fact that for most use cases there is networking gear to be had that is just as good as anything Cisco can provide for a fraction of the cost.
And so Cisco's PR, legal and even political battles against Arista, Huawei and VMware go on.
So far, so normal. That's life in IT. Microsoft's Windows 10 shenanigans or pretty much everything Oracle's ever done prove that this is just the way things are done.
While it is tempting to write Cisco off, it is dangerous to do so. The Cisco of today may be highly dependant on its networking revenues, but it has made key investments that may well ensure it has a long and bright future.
In addition to networking gear, Cisco makes servers. They aren't the most affordable or the sexiest, but they are very popular amongst enterprises. This is because Cisco has been investing in software to make managing these servers (and server/networking combinations) easier.
In theory, Cisco was to integrate Whiptail's Invicta storage and create a compute/storage/networking trifecta. That failed. It cost Cisco a lot of money, a lot of credibility and precious time. Cisco will need to try again with another storage company soon.
If it has learned its lessons from the Whiptail debacle, it might also have an infrastructure endgame machine here soon and be ready to compete with HP and Dell for the "go big or go home" market.
Of course, servers aren't Cisco's only backup plan. Cisco's acquisitions of Metacloud and Pistoncloud gave it Openstack chops that could save it. It's up to Cisco's new CEO to turn things around, though he is not off to an auspicious start.
Three companies, three completely different approaches, but one end goal: to be all things to the enterprise data centre. To build to scale and simply to survive the commoditisation of IT hardware supply.
All three want the high end of the market, but it is too early to tell if there is room for three players in that space. A lot depends on public cloud adoption.
What percentage of workloads will ultimately move away from on-premises data centres and will the ongoing privacy and data sovereignty debates raging between governments around the world result in an increase of regional service providers, or the overwhelming dominance of three big American ones?
In part 3, I'll take a shot at decoding IBM, Lenovo and Supermico. You can find part 1 (Microsoft, Oracle and Amazon) here. ®