Intel's Tablet, IBM's Memory--A Tale Of Two Giants
Intel plays tablet catch-up, while IBM continues to master complexity and profit at it.
No two companies figure in the history of computing like IBM and Intel. Because of each company's laser focus on the business and continued commitment to R&D, among other things, both have weathered the Great Recession better than most rivals, while continuing to define many aspects of computing. But their paths have been far from identical.
IBM continues to court the biggest customers with the biggest challenges, and continues to win with its attitude that difficult problems are rarely easy to solve. By embracing complexity, Big Blue has performed well, one might even say flawlessly--both in terms of its notion of what its market wants and its execution on those customer desires. More on Big Blue's new breakthrough in a bit.
In contrast, Intel has had a rougher time of it during the past few years. The company completely missed the smart phone era and has shown up very late to the tablet game. Granted, both of these are comparatively low margin businesses versus the desktop and server markets where Intel dominates, but by its own reckoning, the Internet needs another server for every 600 smartphones and every 100-plus tablets. That's a lot of devices not using Intel chips for every one that does. Having said all that, Intel remains the undisputed leader in silicon manufacturing. Its processes lead the rest of the industry by a good 18 months or more. So at least technically, there's no reason why Intel shouldn't be able to find a way to participate in these markets.
And, as of Wednesday, Intel's latest Atom chip can indeed be found in a tablet from that household name in personal computing: ViewSonic. Intel says it's the first of many, and while I'm sure it's a fine device, the fact that the first Atom Z670 tablet to hit the market comes from a commodity maker has to be disappointing for Intel. One cool thing about the ViewSonic 10pro: It'll run both Windows 7 and Android 2.3--though the latter runs as an application on the former.
But Intel's not just late to the tablet game. Intel, and everyone else with the possible exception of Apple, underestimated the importance of the tablet for business computing.
InformationWeek's research findings serve to sharpen that point. In 2010, we asked IT pros how various mobile platforms would fare in terms of business productivity during the next two years. Smartphones were predicted by 89% to either increase (somewhat or significantly) in importance, followed by laptops at 53%, netbooks at 38% and tablets at 36%. We asked the same question just a few weeks ago, and while smartphones still topped the list at 82%, tablets had moved up to 79%--with laptops falling to 34% and netbooks to 22%. That's a remarkable movement the likes of which we've never seen before.
It's clear that Intel Atom-based tablets would take years to catch up to and surpass ARM-based devices such as the iPad--if indeed they ever do at all. So what's a chip giant exuding healthy paranoia (as Andy Grove famously said) to do? At least a partial answer for Intel has been to work to reinvigorate the notebook market.
Enter the Ultrabook
Back in late May, Intel announced what it referred to as a thin, light mobile computer called the Ultrabook. The goal: Address many of the shortcomings of the netbook generation while preserving most of the benefits. Ultrabooks are supposed to have much more powerful processors, all-day battery life, turn on instantly, be more secure, and have no moving parts like hard drives or fans. Plus they're less than an inch thick and weigh less than 1.5 pounds. Picture something a lot like the Macbook Air, but probably cheaper. Intel says that it expects at least 40% of notebook users will become Ultrabook users.
Late last week, Intel put some money where its mouth is by creating a $300 million Ultrabook fund that will be used to drive innovation for the system. The upshot for users is that if you were thinking about spending $500 to $700 on a tablet because you like capabilities like light weight, very long battery life, and instant-on, but wish the device came with a keyboard, you will soon have a choice. With clever use of rotatable screens, the Ultrabook could indeed be a better option for some business users who want the best of the notebook and tablet design in a single device.
IBM's Memory Breakthrough
While Intel looks to regain its footing in end user computing by addressing the shortcomings of netbooks and notebooks, IBM is showing off its R&D prowess with a memory technology that will eventually address the shortcomings of flash memory.
The miracle of flash memory is that it retains data even when powered off, and does so without moving parts. On the down side, it's expensive, it's nowhere near as fast as DRAM (though much faster than hard drives,) and the memory cells decay with each write. After about 3,000 writes, the cell can no longer be used.
IBM's new technology, announced in June, is called phase-change memory. It works by changing the chemical state of a compound to produce a variation in the compound's resistance. The change in resistance can be used to store one or more bits of information in each cell. While not as fast as DRAM, PCM can be read and written to at least 100 times faster than Flash memory. But more importantly, IBM says that each cell can be written to at least 10 million times, which makes it nearly as durable as DRAM-- and thousands of times more durable than Flash.
About the only advantage Flash memory has is that it can be packed more densely--that is unless more than one bit is stored in each PCM cell. IBM calls this multi-level phase-change memory, and while a bit of speed is sacrificed (the write process is iterative when storing multiple bits per cell), the result is a device that performs much better than Flash in every way. IBM expects commercial production to begin in about five years.
Don't throw away your venerable 5.25 inch hard drive. You can haul it out and tell your great-grandkids how computers used to rely on platters of spinning rust to store data. Chances are, they might not believe you. Alternatively, you can flip on your Utrabook and just show them a 3-D video of how it used to work.
But as usual there's more going on at IBM Labs than that. The company is also working on what it calls Cognitive Computing technology. The company has produced experimental chips that actually mimic the neurons and synapses of the brain. The recently-shown chips contain 256 neurons and more than 320,000 synapses--most programmable, others self-learning. While that's a far cry from the brain's complexity at this point, IBM's goal is to build a system with 10 billion neurons and trillions of synapses. (Hollywood will have a field day with this one.)
The project is DARPA funded and progressing with help from consortium partners at Columbia, Cornell, UC Merced and the University of Wisconsin, Madison. The goal: Create systems that easily deal with huge amounts of data much as the brain does with input from our own five senses and other systems throughout the body. While IBM's memory technology should be commercialized in five years, computers that are smarter than we are (and that know it) are likely more than a decade away.
Art Wittmann is director of InformationWeek Analytics, a portfolio of decision-support tools and analyst reports. You can write to him at email@example.com.
To find out more about Art Wittmann, please visit his page.
InformationWeek Tech Digest August 03, 2015The networking industry agrees that software-defined networking is the way of the future. So where are all the deployments? We take a look at where SDN is being deployed and what's getting in the way of deployments.