web analytics

Abstract, Jackson Pollock

Technology develops at two speeds.

Most of the time it moves at a smooth pace. There is a constant stream of small, incremental updates, bug fixes and minor changes. Software version numbers tick over by decimal point or less.

Now and then we see a great leap forward. An unexpected new device, app or service emerges almost from nowhere.

Everything changes

Great leaps change everything. They destroy old categories and create new ones. Think iPhone, Google search, tablets or Amazon Web Services[1].

It has been a while since the last great leap forward. We can argue about what it was in the comments if you like[2].

Despite the industry pushing hard in 2015, nothing seismic happened. It was a year of consolidation.

At best we can say some companies laid the seeds of possible future great leaps. While we won’t know what they are until they flower, driverless cars may be one.

This isn’t unusual. Great leaps forward don’t happen every year. Indeed, there are more years without great leaps than with them.


Something that is almost the inverse of a great leap forward happened in 2015. We saw stagnation and worse.

The worse would be Personal computers. These have struggled for years. Sales continue to fall. There has been little noteworthy innovation since Apple introduced the MacBook Air.

Touchscreens were a fizzer. The biggest slump in PC sales coincided with the mainstream arrival of touchscreens.

The problem is with Windows PCs. Apple sales have grown while Windows computer sales show double-digit declines.

Even Apple shows weakness. The new MacBook was a clever attempt to meet a market need. It works well for some users, but sales haven’t caught fire.

Bright spot

One bright spot is advanced hybrid portables like Microsoft’s Surface Pro and the Huawei Matebook introduced last week. Their sales are growing at around 20 percent albeit from a small base.

Hybrid sales may be growing, but they aren’t winning new hearts fast enough to plug the gaps elsewhere in the PC business.

Phones have replaced PCs at the centre of our working and online lives. For years they developed at a clip adding exciting and useful new features each year.

In 2015 that came to a halt. Phones got a little better. Cheap phones got better. But there was no compelling new features to justify faster-than-essential upgrades.

Phone technology has stagnated. No doubt there are cool new ideas bubbling in the labs in Apple, Samsung or Huawei. But, at least for now, we appear to have reached the limit of existing phone formats.

Fixing Windows

Windows 10 is a better effort from Microsoft than the limp Windows 8, but it does little more than put the operating system back on track. It is not a dramatic breakthrough. It is a much-needed fix.

Microsoft scores better elsewhere. Office continues to evolve and spread its reach to new devices. The Surface Pro 4 and the Surface Book are great computers, but neither is a great leap forward.

The most hyped introduction of 2015 was Apple’s Watch. It wasn’t new, a slew of watches were already on sale when it arrived.

It wasn’t special. Sure, many people reading this will disagree and tell me how their Watch is an essential part of their life. It’s almost a year after the Watch first appeared and there is no buzz. Even developers have given up sending out gushing press releases about Watch apps.

Virtual reality continues to generate media copy. Devices may pour out from factories, but little is happening with content. And the content that has appeared is lame.

  1. This, it turns out, is how most life scientists now think evolution works. Smooth changes and occasional big jumps.  ↩
  2. My picks for the most important recent developments are software defined networking and network function virtualisation. While the technologies are big news for service providers they’re invisible to everyday technology customers. That’s hardly the stuff of a great leap forward.  ↩

Twilight of the gods

At heart, Dell and EMC are hardware companies. They face a serious problem: customers have lost interest in hardware.

On October 12 Dell said it will pay US$67 billion for EMC. It’s the biggest technology deal ever and the latest in a never-ending wave of mergers.

Dell may pay too much for EMC’s business. The offer is more than twice the total equity and about 50 percent higher than the value of EMC’s assets. Acquiring companies often overbid to see off rivals.[1]

Most observers expect Dell to sell unwanted parts of EMC when the deal completes. This will lower the final cost of the acquisition.

The elusive benefits of tech consolidation

Big technology company mergers rarely succeed. I can’t think of any that have been a resounding success.

Expect managers to prattle on about synergy as if they know what they are talking about. They rarely do. In most cases the movers behind takeovers are lucky if the new organisation is worth as much as the sum of the parts. Often big tech mergers destroy value.

There is a long list of disappointments. Oracle’s Sun Microsystems acquisition saw a huge destruction of value. At the time there was a lot of talk of synergy.

Then there is HP. It binged on Compaq, EDS and Autonomy — all, allegedly, bringing synergy. There’s little evidence of that.

Maybe that experience prompted HP boss, Meg Whitman, to point out the integration challenges Dell face.. She should know, her predecessors left her a mess.

Dell’s debt risk

Dell faces another risk. It needs to take on a huge debt for the acquisition. The good news is Dell has a track record of paying-down debt. It has already paid back the cost of its 2013 buy-out.

EMC’s shareholders get a profitable exit just as the future starts to look uncertain.

Its operating income last year was US$4 billion, net income was US$2.7 billion. Although debt is cheap at the moment, paying the interest will still be a challenge.

Cisco shareholders must be praying someone else with deep pockets is thinking of a similar move.

Cloudy forecast

Dell’s EMC buy underlines how cloud computing has hurt enterprise technology companies.

Cloud is the biggest change since networks of small computers replaced mainframes in the early 1990s. Unlike the earlier waves of creative destruction this isn’t about replacing a hardware generation. It’s about the move from hardware to services.

EMC has a toehold in the data centre business. It sells storage hardware and virtualisation software to help spread loads between servers. This is a company that until recently had double-digit growth. Today EMC is growing at a slower pace.

Dell diversifies with EMC

Dell has moved into corporate computing. It has made big strides with storage and servers. Yet it still depends on PCs for most of its income. That market is in free-fall.

The server business is no better. Only IBM and SAP have much in the way of mainframe business. Players like Unisys are now consigned to mainframe niches. And that sector is unlikely to grow.

The smaller servers that have dominated business computing for the last twenty years are also on the way out.

While the shift to cloud computing threatens Dell and EMC, it also threatens all the other business technology giants of the last 25 years.

Microsoft is an exception. It got into the cloud ahead of the trend and is now the main competitor to AWS which came from nowhere to dominate enterprise computing.


Away from the cloud, hardware makers are now churning out basic, commodity computers which can be programmed to act as servers, routers or storage devices. Hence Software-defined networking, software-defined storage and even software-defined data centres.

This trend also reduces the once valuable integration revenue — everything plays together better now.

Dell’s strength lies in building low-cost, undifferentiated hardware. But even its commodity computers are too expensive for the big cloud companies who build their own kit or outsource the job to white-box contractors.

Now Dell is banking on selling low-cost hardware to enterprises so they can build thier own private clouds. It may also hope to juggle the economics of building commodity hardware to win back business from the cloud companies.


Many commentators view the deal as a sign of the triumph of new cloud players over old enterprise technology companies.

There is something in this, but it’s not straightforward. All the old companies have made huge investments in cloud technology. They just haven’t moved fast enough, or confidently enough towards the new way of running technology.

Unlike AWS, they face a problem, if they race too fast to the cloud they’ll replace their existing high margin business with a low margin alternative. Only Microsoft has avoided this fate.

Which brings us back to Dell buying EMC. It isn’t a triumph or a new beginning, the deal is a step in the death throes of an era. What we are seeing is the twilight of the technology gods.

  1. Once a takeover is in play, it can often be just as important, maybe more important, to keep the target out of rival’s hands. Much of this is just something that goes on inside manager’s heads. I can’t remember many large scale acquisitions which ended up causing serious damage to rivals.  ↩

The image at the top of the story is Siegfried and the Twilight of the Gods by Arthur Rackham.


Mainframe maker Unisys has woven a new enterprise computing system it describes as a “fabric-based architecture”.

Branded as Forward, Unisys’ new architecture stitches together teams of Intel Xeon processors with a fast backplane and embroidered with the company’s own s-Par advanced secure partitioning.

Forward aims to help large businesses and government departments move from proprietary Unix systems.

Like a mainframe, not a mainframe

Australia-based Forward program manager Vic Herring says the new architecture is designed to cut through the problems organisations face when moving to new architectures. He says Forward brings mainframe levels of security and availability.

In other words Forward plays to Unisys’ strengths in building mission critical systems for organisations with big workloads. It ticks all the important 2013 enterprise computing boxes: Forward can be used for cloud computing, big data and run large ERP applications like SAP.

Virtual and then some

Herring says Forward is way ahead of other virtualisation systems. Instead of conventional virtual machines Forward provides a series of secure partitions tightly coupled to the Xeon processors. Each partition handles a specific workload and has its own dedicated resources: processor or processors, memory and input-output channels.

He says these resources don’t bounce up and down when other jobs stretch resources on other partitions – the way things can work with conventional virtual machines. He didn’t quite say that. Herring talked about resource contention and latency.

One of the features of Unisys’ fabric computing approach is that communications between each secure partition takes place at the same speed as memory transfers. Data doesn’t have to queue to get aboard the bus.

Herring says Forward uses software driven architecture. This means many things that would normally be done by hardware are software controlled. So, connecting two devices within the fabric no longer requires Ethernet cable but is done with a software command. Even devices like routers in effect become software routines. In this respect the technology resembles software defined networks.

Unisys does stealth security

Security is important to Unisys customers who often deal with large volumes of sensitive data. Each of the Forward partitions is secure in the sense that each partition is fully containerised. So if a job in one partition falls over, it doesn’t affect other partitions.

The new system also can be used with Unisys Stealth security software – which the company also sells as a standalone offering. Stealth uses cloaking, which means only authorised systems know a protected system is on the network. It also has encryption. In effect systems protected with Stealth are invisible.

Not cheap, but cheaper

Herring is frank about his company’s reputation. He says: “I know in the past Unisys hasn’t been the cheapest on the market”.

While it still isn’t playing in the bargain basement, he says Forward uses off-the-shelf components to deliver the enterprise computing power at the fraction of the cost of alternatives. Computer makers often throw numbers around, but Herring’s are impressive. He says Forward cuts capital costs by 50 percent or better and reduces the total cost of ownership by 40 percent.

Software defined networks are set to do for networking what virtualisation has done for servers. There’s a great story explaining them in the Sydney Morning Herald’s ITPro section.

Virtual machines turn a single physical server into what appears as several separate computers. SDN separates network hardware from the controlling software. This means switches and routers can be reconfigured and orchestrated just like virtual servers.

Managing a network becomes flexible and dynamic.

The Next Big Thing

When companies first virtualised servers, it was mainly about consolidating resources. Quickly they found other benefits and the flexibility meant entire systems could be reconfigured at the drop of a hat. That’s probably what will happen to networks. You can expect rapid changes in storage technology as SDN gains acceptance.

Although few New Zealand businesses run their own networks on a scale that can benefit from SDN at the moment, you can expect telcos, cloud computing companies and other service providers to use them in coming months.