More like this

Data Centre

Supermicro's ability to enable should worry IBM and Lenovo

With customer backing, the tech minnow is trying something rather different

Data scientist image via Shutterstock

Sysadmin's 2015 review part 3 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 constantly changes, and so here's one sysadmin's final view of the industry's movers and shakers.

In part one (link here) I took a look at Amazon, Oracle and Microsoft. In part two (link here) I tried to decode Cisco, HPE and Dell. Here in part three, I am going to look at IBM, Lenovo and Supermicro. IBM is returning to its roots as a maker of the bizarre and the esoteric, while Lenovo is trying to become IBM from the 90s. Supermicro, meanwhile, is being driven by its customers to become something altogether different.

IBM

Of all the large tech companies, IBM is the one I find most interesting. If there is a consistent strategy governing its efforts, I have been unable to discern it.

More than any other company in tech, IBM seems beholden to Wall Street. Everything IBM does seems tied to its share price. Quarterly return concerns and share buybacks dominate public discussions and news articles.

IBM has been artificially supporting its share price by selling itself off to Lenovo a piece at a time but eventually it is going to run out of business units to sell. IBM simply isn't the behemoth it once was, and that ever-important share price has been on a downward trend since March of 2013 for exactly that reason.

If you want to be a gigantic company, then you actually need revenue. This means you need business units that generate money. IBM hasn't seemed to clue into this yet.

IBM relies heavily on providing IT services to enterprises. Unfortunately, IBM is facing increasingly competent competition from virtually every other tech giant on the one side, and on the other, growing antipathy towards the quality of its own offerings from its own customers.

IT services is a fickle business to be in and because of this it is best to have services be a secondary revenue source than a primary one. IT services can generate demand for other technologies provided during booms and retain customers during busts ... but you need those other technologies in the first place.

IBM's SoftLayer cloud offering may one day be such a technology. Today it simply isn't enough. IBM is one of the largest public cloud offerings in the world, but compared with Amazon's AWS or Microsoft's Azure it might as well not even exist.

Like pretty much everything IBM does these days the primary selling point of SoftLayer is that IBM can automate technologies that other vendors aren't willing to. Power-based servers, large enterprise storage arrays, you name it. The more niche the request the more likely only SoftLayer can make it happen as a public cloud service.

The flip side to this is that IBM is essentially serving as a pilot project for everyone else. IBM determines what people are willing to pay for as a public cloud service and then Amazon builds the ones that see decent traction into AWS.

IBM has always focused on the niche and the mind-bogglingly expensive end of the large enterprise. While everyone else is more than willing to leave it serving the niche demands, Dell, HPE and Cisco are targeting the very same mind-bogglingly expensive end of the large enterprise for all they're worth.

Against this backdrop, IBM still does research. Because it's IBM it does research into the bizarre and the esoteric. Much of it doesn't pan out, but every now and again it comes together.

Worth nothing is IBM's new Cognitive Business Solutions unit. Two thousand analysts and big data scientists pimping out Watson's big sexy digital brains for all they're worth. Decades of research have culminated in machine learning software that IBM can make do what nobody else can.

IBM is also doing research into everything from silicon photonics to quantum computing. It does nanotechnology research in rooms so quiet they will make you vomit. And don't forget that it still makes mainframes. Mainframes that run workloads no other computers can run, which is why they're still around.

What IBM can't seem to do is commercialise this research. IBM is stuck in the past, a past that doesn't seem to have moved very far from Thomas Watson's (president of IBM in 1943) hilarious (and possibly not actually existing) prediction that "there is a world market for maybe five computers".

That attitude — that IBM builds things which appeal to a very small number of clients and then charges the moon for them — just doesn't work any more. It isn't that the research should stop, it is that what IBM builds needs to be brought to the mass market quickly and cheaply.

We no longer operate in a world where high technology and manufacturing are the province of only a few organisations. We live in the world of makers, of 3D printers and laser sintering, and of semiconductor fabrication capacity that can be had for a song.

IBM has consistently been granted more patents every year than any other company in the US. What IBM needs now is to find someone willing to actually put them to use.

Lenovo

If IBM is a research powerhouse catering to the wealthy, Lenovo is the megamall of the tech industry: it has something for everyone and the more mass the market the happier it is.

Lenovo is something of a modern curiosity in the tech industry. While everyone is rushing out to try new business models, Lenovo is instead trying on IBM's old one.

Amazingly, Lenovo seem to have the unique ability (at least within the tech industry) to learn from the mistakes of others. For example, Lenovo wants to appeal to the mass market. It wants to sell you desktops and notebooks, smartphones and servers, while maintaining quality.

So Lenovo is trying to make good gear. As one of its largest markets is China it is also trying to make inexpensive gear, but it seem dedicated to actively working towards a reasoned balance between quality and affordability. It even – shock and horror – offers tiers of equipment that occupy different balance points, each appealing to different market segments.

In addition to this, Lenovo is building out a global services arm, partnering with traditional enterprise hardware and software providers and even in-demand startups such as Nutanix and SimpliVity.

What has been missed by most is Lenovo's increasing commitment to research. Lenovo has been sponsoring education initiatives around the globe, and has been building research parks, funding startups and – above all else – looking for new markets for its technology.

The IBM of old did exactly this. It was an organisation dedicated to finding new ways to incorporate technology into every aspect of our lives. Seen and unseen.

While Lenovo's research focus seems to be largely emerging markets the needs of these emerging markets are driving demand for previously unthought-of applications of new and existing technologies. Lenovo's acquisition of Motorola is helping it here, as small, low-power devices for emerging markets are slowly becoming a Lenovo speciality.

Today, Lenovo's research spend as a percentage of revenue is still significantly behind powerhouses such as IBM, Intel and so on. It has been increasing and looks set to keep increasing with time. Whether Lenovo follows through or not is, of course, open to speculation.

Having reshaped itself in IBM's image circa the 1990s, the question remains if it can avoid IBM's mistakes from the 2000s.

Supermicro

When putting together my list of companies to care about I had a half page of notes on all vendors in this series, but one. Included amongst the notes was Supermicro, excluded was Cisco. When a colleague of mine pointed this out, he said to me that this was probably the single most important thing I could possibly say about Supermicro.

Supermicro is a fraction of the size of the other companies in this series. Mentioning it alongside Microsoft, Oracle, Amazon, Dell, HPE, Cisco, IBM and Lenovo is to include Supermicro in rare company. Still, the fact remains that when I was picking out companies that are real threats in the IT landscape, Supermicro came to mind and Cisco didn't.

Supermicro makes damned fine computers. It makes great switches and storage, too. With every other vendor you need to find the offering that is the closest fit to your needs. With Supermicro it will have something that exactly fits your needs. If it doesn't, then it'll make it.

This, right here, is why it is a threat.

Every now and again someone posts a news story about how some company bought a lot of Supermicro gear. That's neat and all, but this is now a regular enough occurrence that it's not really news anymore.

What makes Supermicro a threat is not its selling en masse to end customers. It is its position as an enabler. Supermicro is the secret weapon of Silicon Valley startups. Nutanix would not be where it is without Supermicro servers. Heck, more than half the storage industry and almost every backup applicant every use Supermicro gear.

In addition to the hardware, Supermicro has a global enterprise support offering that has had years to work out the bugs. Supermicro makes it possible for a startup with no more than a handful of people and some software to start shipping enterprise class gear, with global support.

Add in partnerships with a few managed service providers to handle installation and Tier 1 support issues and you're off to the races. The great big vertical integration plays of HPE, Dell and Cisco will not let them freeze out the competition so long as the competition has access to Supermicro.

Supermicro has become a two-billion-dollar a year company and doesn't look set to slow down. It's done this by enabling startups to make life difficult for the established players. That makes it critical: today, tomorrow and for the foreseeable future. ®

Sponsored: Accelerated Computing and the Democratization of Supercomputing