Intel’s Optane memory is the first “new” storage technology to appear on the consumer and enterprise market since the earliest iterations of SLC NAND flash started showing up in enterprise solutions a decade ago. While the products have, thus far, been limited to a small slice of the consumer and enterprise markets, Intel intends to continue investing in the technology and to bring more products to market, PCWorld reports. Long term, the company wants to push all local storage out of the market altogether.
Intel is also pushing hard for 5G standard development; its XMM 8060 modem is supposed to be one of the first 5G radios to market when it arrives in mid-2019 or early 2020, and the XMM 7560 is its first CDMA-capable modem, which could help the company take over Apple’s iPhone business.
In short, Intel is making some major plays into cellular connectivity, and its concept for cloud-connected PCs is a direct consequence.
Intel’s basic idea is that consumers would have an Intel Optane drive as a cache drive (how large is not specified), with all other data being streamed out of the cloud. Now, in theory, this should work reasonably well, provided the Optane drive is large enough to keep critical local files handy, and PCWorld points out that services like Microsoft OneDrive currently show files as being stored locally that are, in some cases, actually stored in the cloud. And while the battle between thin client and fat client processing has literally been going on since computers were first powerful enough to support remote terminals, we’ve never seen the entire industry embrace the kind of cloud-connected PC that Intel now thinks it can push.
And for good reason. While this concept could absolutely work for some specific use-cases like education and businesses where local storage needs are small and security is paramount, it’s a terrible idea for the general consumer market.
First, 5G refers to a cellular standard, not a type of Wi-Fi, and cellular data is anything but cheap. The more of the day-to-day files someone uses that get shoved into the cloud, the more data they’re going to chew through on an ongoing basis. If you’ve ever tried to use the cellular or Wi-Fi networks in congested areas or peak usage times, you’re aware performance in these situations can crater and stay that way for long periods of time. I’ve stayed in gorgeous hotels with the Wi-Fi performance you’d expect from a router buried under 30 feet of concrete.
It’s not even accurate to say that connectivity in these areas is improving. Sometimes, it isn’t. I moved to upstate NY in 2011, and while I’m scarcely a one-man Michelin guide, the same places where I noticed having terrible cell service in 2011 have terrible cell service in 2017. Nor is there any reason to think that 5G, which requires more base stations and towers, thanks to its reliance on shorter millimeter wave networks, is going to lead to some kind of investment revolution courtesy of AT&T. Especially not with the FCC bending over backwards to find reasons to declare internet broadband a solved problem in the first place.
If you’re a gamer who cares about visual fidelity, the only thing cloud services gets you is a substandard title yoked to video compression and input lag. If you’re editing video, or working on a 3D rendering project, streaming in resources is an obvious step backwards compared with having them stored locally. Intel pushed back at this argument by framing it in terms of hard drives, pointing out that the performance impact of putting an HDD in the cloud is approximately equal to the performance of a SATA-based hard drive if connected via 5G. But even if we take the company at its word, we’re ignoring the fact that people dumped hard drives and went solid-state for a reason.
Are there uses for this type of technology? Sure. Always have been. And people have been installing thin terminals ever since thin terminals became A Thing. But that doesn’t make them a great solution for the market as a whole, and even if we see systems like this appeal to a larger range of customers, they’re not going to address the needs of anyone who actually wants to do serious work with a system. They would, however, help Intel sell a lot of modems. Once you know that, you know pretty much everything else about why this is suddenly such an interesting idea to Chipzilla.