The Internet of Things is creating an enormous amount of data. To manage, access, and make use of this data, digital storage becomes a critical factor. We spoke with the authority on digital storage, Dr. Tom Coughlin, to get his insight into how the industry is reacting and where innovations are leading.
Dr. Tom Coughlin has worked for over 35 years in the data storage industry at numerous companies. He has over 100 publications and six patents and publishes reports on digital storage technology and applications. Tom is active with the IEEE, SMPTE, IDEMA, and other professional organizations. He is the founder and organizer of the Annual Storage Visions Conference.
He is also a partner to the annual Consumer Electronics Show as well as the Creative Storage Conference. Coughlin Associates, of which he is Founder and President, provides market and technology analysis as well as data storage technical and market consulting. Tom is also general chairman of the annual Flash Memory Summits.
Predictions estimate 20 billion devices will be connected to the Internet by 2020. Is the storage industry prepared for the onslaught of data?
I think they’re trying to get prepared. There’s going to be an awful lot of input of information from all these sensors and smart devices that communicate with each other and also communicate to a broader audience.
As I see it, there are some interesting new models for a sort of ad-hoc local or edge computing and edge storage model as well as more aggregated cloud storage. A lot of the stuff people are looking at for the IoT would put information storage in the cloud. However, it may make more sense for local storage and processing of data to minimize the long-term aggregated data in the cloud and to control the bandwidth requirements. These bandwidth requirements will be huge if the storage and analysis of raw IoT data is done in data centers. It may make a lot of sense to do the initial analysis of the data at the local level and only send processed results to the cloud. That could also provide ways to preserve privacy if this is consumer data – privacy and security of that data – especially if the data can be shared anonymously versus having precise metadata tags that tell you who and when every piece of data is associated with, which would be available in the raw data.
What types of technology do you see that are available today to put storage into the IoT devices, and is there anything new or cutting edge that you’ve seen?
Yeah, there’s some interesting possibilities. Of course, a lot of the IoT devices, especially the smaller ones, tend to use solid state storage, and thus flash is commonly used. There are other possibilities, but flash is the least expensive. It’s fast enough for many of those applications, and it’s relatively low power, particularly since it doesn’t need to be refreshed like say DRAM or SRAM does.
“I think that we’ll be seeing a proliferation of devices with flash for storage”
I think that we’ll be seeing a proliferation of devices with flash for storage, and then I think there may be some other new solid-state storage technologies that will be emerging that could also play some roles.
Intel Micron had talked about their 3D XPoint Technology. Which I think for the next few years they’re targeting more towards server application, but they’ve also indicated a possibility for gaming and other purposes as well. This technology is going to be a bit more expensive than flash, less expensive than DRAM, with speeds that are approaching that of DRAM.
Also, there’re some other technologies like Magnetic Random Access Memories (MRAM) – that could also be used for caching and buffering and used like that. MRAM and especially STT-MRAM uses very low power and is a very fast technology. I think flash will probably provide the lion’s share of storage in a lot of these applications unless it’s a larger device; then you might be able to have hard drive storage. But there may also be new, emerging solid state technologies that could be playing a role as we get into the next decade.
What are the advantages and disadvantages for hard disk storage versus using flash?
Hard disk drives are advantageous if you have a lot of stuff you have to store because they can store that less expensively per capacity, depending on how much you’re using it. Just in terms of the raw purchasing costs, the dollar per gigabyte today is less on a hard drive than saving the same data on flash.
Flash may actually be less expensive than hard disk drives to provide some level of storage capacity since the HDD may have much more storage capacity than the required storage level. Today that is probably about 64 GB but in the near future it will be 128 GB and then 256 GB. If you only need a bit of storage, and you can get that for less than say $40, then you may be able to get the storage capacity you need with flash memory for less than an HDD.
You mentioned gaming. The gaming industry has typically been an early adopter of new storage systems. Which vertical markets do you believe are best poised to capture new storage opportunities fueled by the Internet of Things?
Speaking of gaming, I think both for visualization and also gaming applications, virtual reality seems to be very popular, and I see it growing. That actually could be one way in which to visualize content from sensor networks and that sort of thing. If I can create a visualization and see that in a 3D display, it may give me a sort of geometrical sense of the data. I think that would be a popular way of doing it.
“If I can create a visualization and see that in a 3D display, it may give me a sort of geometrical sense of the data.”
Also sensors, networks, and Internet, these things may play a role in future games, in particular in these virtual reality-type environments as well since they’d give you new ways in which to get a more immersive experience, as long as you don’t run into things.
Have you seen anyone in the gaming space doing things you find interesting?
I’ve seen some things that people are starting to work on in the virtual reality space. I’m pretty sure that the reason there’s been investments, say by Facebook and others in this space, is for finding new ways of representing information in ways that people can address, and visual is a very good way of representing stuff.
If you do it in three dimensions, it gives you new ways of projecting and using that data. So I think we’re going to see some virtual reality-type experience being used to augment our data analysis and visualization technologies and approaches. I think it could add new kinds of gaming and simulation type of characteristics as well that could be useful.
In regards to hosting companies, do you think the hosting companies will need to bulk up to handle the data from the IoT?
Even if we have to use some kind of processing and analysis on the edge, just the sheer number– because the price of processing is going down, as well as storage, and we’re getting more or less ubiquitous networking, we’re going to have more things that are going to have networking and storage and processing built into them. That means they’ll generate information.
A lot of that information, even if it’s analyzed and processed, will end up in some form in the web. A lot of that’s going to be in data centers and therefore in the cloud. So this is going to be a big driver for cloud growth, in my opinion, just with the sheer number of things generating data, ultimately some of this data will make it into the cloud.
Related to that, have you looked into AWS Cloud IoT services? And do you see them as a leader or do you see any significant competitors?
I think we’re going to see a lot of people in this field, whether it be Google, or Facebook, or others as well. One of the interesting things that I see is that we have a lot of connected devices associated with us as individuals, essentially that is an expansion of our social network. We’re creating electronic prostheses, digital appendages that become part of our social network and part of the ways in which we interact with others.
“We’re creating electronic prostheses, digital appendages that become part of our social network and part of the ways in which we interact with others.”
So I think we’re going to see a lot of different services. AWS is one, but I think we’re going to see an expansion of those kind of services that are available from a lot of different sources for a lot of different applications.
There’s a lot of focus right now is on the industrial side of it because there are some pretty immediate paybacks. But ultimately I think we’re going to see a lot of consumer applications for the IoT, as well. All of the smart devices that people are wearing, carrying with them nowadays, are going to be contributing to that.
What do you think the market looks like down the road? Do you think we will see specialty companies serving niches like consumer versus industrial? Or is there going to be a dominate player like AWS serving everyone?
For a lot of the public Cloud applications, there’s an economic value that you get from the amount of scale that they have. So I think that definitely can be an advantage.
But as in any market you probably are going to have different niches with certain specialized areas that don’t fit the general mold. Therefore, there may a be a lot of different types of opportunities. Anything like this, even if you’re working with commodity and community-accessible assets, you may still get some differentiation. At least at the beginning before it gets totally commoditized. I think we’re just in the scale where that sort of thing is going to be happening. I don’t think it’s one winner take all for a little while at least.
With the IoT generating an enormous amount of new data, do you think legislators are going to get involved with new laws and regulations?
I think that there’s going to be issues and things that are going to have to be resolved concerning privacy and security of personal information. I definitely expect those kind of things.
I think one of the ways to deal with that is if the actual exchange of information, personal information that may contain personal content, is done anonymously so that you can’t just be tracked with everything you do just because you happen to have sensors that you carry with you.
How we can secure data on the IoT devices, and is this the responsibility of the developers or are there standards that are beginning to emerge or will emerge?
I think that eventually it will probably be the standard, but right now, it’s going to limit the growth in this technology, unless it’s addressed, and I don’t think it is fully addressed yet.
I think there’re plenty of indications, from all of the things you read about people being able to take over your Chromecast and access your network to people being able to hack cars and make them do things they want them to do, we have a lot to learn with regard to security in our connected devices. So I think it’s something that needs to be addressed.
I believe there are ways in which you can do that, one of which might be to have an initial focus on these localized edge networks, and making them secure and having access to them to be moderated and controlled. That’s where, for instance, a choice of some information which you want to share anonymously versus have it be identifiable to you that we need a capability of doing that kind of thing before I think a lot of people will be comfortable with this.
I think we can create technologies that may collect and analyze the data locally and then secure that data, and access to that data, to the outside environment. That may be one approach. I call this concept having a guardian angel that runs on your own equipment that then can interact in something called the invisible colleges of the guardian angels to be able to share and aggregate anonymized information to create composite information, for instance. I wrote a white paper on that awhile back with some other folks, and I’ve done some talks on that in the not too distant past.
Is there anything else happening in the storage industry in regards to the IoT that you find interesting, or do you have any predictions of what’s to come?
I think it’s going to cause enormous changes in terms of how what we store, where we store it, and how we store it. I think it’s going to help drive the acceleration of the process of bringing data and processing closer together because there will be immediate value in doing that. If you have very recent data from a sensor network, the immediate process and analysis of that could have enormous value, so I think it’s going to be driving a fundamental transformation of computer architectures.
“I think it’s going to cause enormous changes in terms of how what we store, where we store it, and how we store it.”
That could include the replacement of volatile and nonvolatile memory technologies and also with new types of architectures that could, for instance, put processing power on the storage device themselves. There are people talking about doing that with SSDs – Samsung has their Storage Intelligence 2.0 – to other types of what’s generally called in-situ processing where the processing device and the storage device are quite intimate.
Ultimately, it may lead to proliferation and development of new technologies that can even more fully integrate memory and processing capabilities. I think all these will be very important for dealing with the big data problem that collection of all this Internet of Things will create.
The Internet of Things series is brought to you by Treeline Interactive and is authored by Tim Homuth and Joe Austin. Our objective is to provide you with information about the rapidly evolving field of the Internet of Things by bringing you interviews and insight from industry leaders. If you have any questions, insight, or suggestions for articles please email us at IoT@treelineinteractive.com.