Life is about cycles. When I think about the cloud, for some reason the Chinese yin-yang symbol pops into my head. And when I consider today’s billions and tomorrow’s tens of billions of cloud-connected mobile devices, inevitable cycles of business and IT centralization and decentralization come to mind.
We are obviously going through a heady time in relation to the cloud and cloud computing services. However, at the same time that we are appropriately amazed at the growth of the cloud and eagerly welcome services such as iCloud, at the same time that we cheer the arrival of new Samsung Galaxy models, the latest HTC entry or the arrival of the iPhone 4S, we forget to look both backward and forward at the evolution of mobile devices and networks. This is especially important since some of the worldview of the cloud seems to assume that the devices on the edge, particularly mobile devices, don’t need to be too smart. And although we believe these new devices to be smart today, we need to begin to focus on how staggeringly powerful and capable devices will be in the future.
A quick look at the cost of storage
I have three data points to use as illustrations, all three of which unfortunately date me as having been around a while. The first mobile business PC I ever had access to was a 1983 Kaypro 10, a CPM-based machine with a 2.5 MHz processor (yes, MHz) and a 10MB hard drive. A machine that promptly crashed due to hardware, or — more likely — user error.
Skipping ahead to 1993, I was living in Europe and needed a new hard drive for the Dell (s DELL) desktop I had brought with me for my expatriate assignment. I found a mail-order hard drive (from the back of PC Magazine), a bargain at $700 for 250 MB.
The third data point was in 1999, when I bought my first digital camera (2 megapixels) and bought my first CompactFlash card, which cost $80 for 16 MB — my first truly “mobile” storage.
Fast-forward to today, as we read the plethora of announcements on the latest smartphones with gigahertz plus dual-core processors and1GB and more RAM, utilizing small expandable SDHC storage that costs barely more than $1 per GB. Or news about larger solid-state drives (SSD) for notebooks at about $700 for 512 GB or hard drives that cost $40 a terabyte. Sometimes we don’t think through the future implications.
So let’s think it through, but let’s not be backward about it. Yes, solid-state media has gone from $4800 per GB to about $1 per GB in 12 years. And spinning media hard drives have gone from $2,800,000 per Terabyte in 1993 to $40 per terabyte today (and yes, I know there are probably other ways of looking at this, but it’s just an illustration).
The 16 MB solid-state card I bought in 1999 has increased in storage size by 2,000 times in 12 years yet the cost has remained roughly the same (actually, if you factor in inflation, the cost is lower). The 250 MB hard drive I purchased for $700 in 1992 has increased in storage size by 2,000 times for the 512 GB SSD, or 12,000 times for the 3 TB hard drive that is one-sixteenth the cost in 2011 dollars!
The cloud is going to need some help
So what does this have to do with the cloud? A lot. Elements of the “cloud is all” ecosystem view in relation to mobile devices is that we all will have ubiquitous, unlimited, low-cost connections from our “really really really smart phones” of the future. And this view often does not take into account the nature and realities of the mobile network or the rapid progression of device capability. And our wireless networks won’t be able to handle this on their own. A few years ago, I wrote a piece for GigaOM on problems with persistent wireless apps, and a piece last year on Verizon’s potential of overpromising of LTE.
I love LTE (even with carrying extra batteries), HSPA+ and all my cool devices. However, available wide area bandwidth defines the wide area network capacity available to any individual user on a loaded wireless network. This will always be problematic in relation to the power of the cloud. The problem will be how we get massive volumes of increasingly broad and pervasive forms of content from our devices to the cloud, and the cloud is going to need some help. A key part of the answer will be to increasingly leverage the power of the mobile device, the storage of the mobile device, and the power of the applications in our hands. This will not be a “nice to have,” it will be a necessity. Amazon gets this concept with the introduction of Amazon Silk. Google gets it with the recent introduction of Google Offline Mail (disclosure: 12 years ago, I was GM of Qualcomm’s Eudora Email Group, so Google’s announcement is sort of “what was old is new again” to me).
The power of the cloud needs to leverage the power of the device
So, a thought experiment. It’s 2020 or 2025. Our N-Core, multi-gigahertz processor handheld device(s) have terabytes (1000 GB) to tens of terabytes of local storage. All the music anybody might want to hear can be stored in a few hundred gigabytes, or at most few terabytes (even uncompressed). Hundreds of full-length movies aligned with an individual’s preferences are stored locally for a few more terabytes, ready to be unlocked and viewed at some “to be determined” cost. For gamers, all the hottest 3D games are already resident on the handset ready to be purchased and streamed wirelessly to the nearest display. Hundreds of gigabytes of personal photos and videos are at your fingertips. The print collection of the Library of Congress, not that we would need it, would be another 10 terabytes. How about having all the relevant content of our favorite websites and social media also resident on the device? Probably not as much heavy lifting as the Library of Congress!
We’ve solved lots of problems in wireless in the last several decades, but for wide area networks, we are getting closer to Shannon’s limit, which defines just how efficient a wide area network can be for a given amount of spectrum. So how will these hundreds of gigabytes or terabytes of data be delivered to the device(s)? Probably from fat local area wired or local wireless network pipes. Which by definition starts making us think about parsing content between real time delivery leveraging limited and costly wide area capacity vs. preloaded/background loaded/resident content delivered on local networks with less costly economics per gigabyte delivered.
Will it be the carriers, Amazon (s AMZN), Netflix, Apple (s AAPL) or new entrants doing the preloading or side-loading? Or if you don’t believe there will be tens of terabytes in mobile devices — maybe only a few terabytes — what if those devices can talk to one another in a peer-to-peer or mesh network? What’s the aggregate power and capability of billions of these things, especially if there will be ways for them to work with and talk to one another both alone and in conjunction with cloud-based services? There’s probably money to be made here in the next 10-15 years!
The moral of the story is to recognize that the nature of the cloud is the nature of all processes of centralization and decentralization, and in my almost three decades in the tech world there have been several in the IT space. The realities and costs of the cloud in relation to wireless services and devices, especially when those devices will be limited by the nature of wide area wireless, and those pesky things called radios, will be a challenge. However, when that power is appropriately aligned with the massive processing power and storage that will be in our hands tomorrow, it may well shape a future which will be beyond the cloud.
Jeff Belk is Managing Director of ICT168 Capital, LLC, investing and working with wireless firms globally. He spent almost 14 years at Qualcomm, in roles including SVP, Global Marketing, and SVP, Strategy and Market Development.