GLGi: The Future of the Data Storage Market
How Far will Flash Go

Thomas Coughlin

President, Coughlin Associates

April 10, 2007
New York

Thomas Coughlin: Hello, this is Tom Coughlin, Coughlin Associates. I'm here to give a presentation as a GLG leader. I'm going to be talking about digital storage topics; it's billed as The Future of Data Storage Market: How Far Will Flash Go?

But I'm actually going to be talking about data storage in general, including flash, hard disk drives, optical storage, and also, in particular, I have a focus on the impact of networking and sharing trends and its role in the expansion of digital storage.

So, we'll start at slide four here. My talk is called Personal Content and Home Network Storage: The Perfect Storm.

Slide six talks a little bit about me, Tom Coughlin, Coughlin Associates. My website is TomCoughlin.com. I've had over 25 years of experience in the data storage industry at numerous companies working on hard disk drives, tape, and, for those of you who are old enough, floppy disk drives. I worked in engineering, engineering management, and senior management positions at different companies. And, since 1999, I've had my own company.

There are three primary things that I do at my company. One is consulting - that includes both technical as well as market, due diligence, and other consulting. Also, I organize some conferences: one called Storage Visions, which focuses on data storage and entertainment creation distribution, and reception, which is a partner of the Consumer Electronics Show. There is a new conference called Creative Storage that I'm doing in May.

I also write reports and articles on digital storage and its applications, particularly related to entertainment and consumer applications. There's a listing on this page six of some of the reports I've put out within the last year.

So, an outline on page seven. I'm going to be talking about the rapid development of consumer applications, creating demand for many types of digital storage. We'll talk about new storage technologies and development of flash, hard disk drives, and optical storage, as well as direct-connect storage devices. In particular information from the 2007 Storage Visions conference, as well as Consumer Electronics Show.

We'll talk about how networking will change the role of management of storage in the home, new rules and guidelines for digital storage design, and talk about how consumers will increasingly dominate the total digital storage content creation. And also, distribution and how greater demand will be driven by inexpensive storage, which also will create new needs in the market.

Slide eight: 2007 CES and Storage Vision products that were shown there. In the upper left-hand side, we see what I think is a very important product. It's put out by LG. It's a multi-blue laser format reader that reads both Blu-Ray and the HD-DVD format, blue laser optical disks.

I think it's very important; devices like this will make it possible to buy one format and still be able to read it even if that format does not survive the format work. So, I think it's an enabler for this market.

Next, to the right, is shown an example of the first introduced holographic storage disks with 30 GB capacity from InPhase in Longwood, Colorado. To the right of that - we're moving clockwise here - Sygate's Free Agent family of drives, including desktop as well as mobile applications.

Then, moving below that, SanDisk slide from the CES show, talking about different products all the way from multimedia players to solid state drives to Vista Ready Boost flash memory devices, and then some new concepts that they showed.

Next to that is a Samsung 32 GB SolidState drive and to the far left and below there, is Hitachi Global Storage Technologies CinemaStar hard disk drive, which was introduced at Storage Visions and at the CES show.

Basically, moving on to slide nine, the CES show and also the reason why I started the Storage Visions conference, involves a sea of storage devices. You find them everywhere. You see flocks of flash memory, herds of hard disk drives, and oodles of optical storage devices out there.

This first slide - slide nine - shows, in this case, a compact flash, an older format flash. But you can see that there are many, many, many different factors, many different capacities out there. There were a lot of different types of flash memories on display at the conference.

The next slide, ten, shows USB drives, and these are flash-based disk drives. To the upper left, is an example of SanDisk's USB ready boost drive for Vista. The purpose of this product is to increase, if you will, the read-cache capacity for computers which have a gigabyte or less of RAM, making them capable of using with the Vista operating system.

Also, flash memory devices are becoming so ubiquitous that we are starting to see developments, which are driven by fashion or design things. Since people may, in fact, carry these with them, or have them even at a chain hanging from their neck or things like that.

We're seeing things like gold-plated USB drives, platinum-plated various designs all the way from very elegant-looking to very silly. For instance, an example of the lighter category are USB drives that are made to look like pieces of sushi.

Moving on to slide eleven, also on the flash memory, we see in the upper left hand side, a slide from the CES show, showing SolidState drive projections by SanDisk. Basically, the flash memory guy's been talking about in the ultra-light notebook market, and in notebook market in general, that by 2010 or so, they're thinking they might be able to get 20% of that total market.

My own feeling is, I will explain a little later, is that without some additional work, I think that is probably not achievable. It may be closer to 10%, but there are paths leading to higher percentage penetration. But probably assuming that it's not strictly a flash memory SolidState drive solution.

In the right-hand side we see again the SanDisk slide showing here connectivity, multimedia devices, phones, as well as video music players, solid-state drives, desktop-ready booster, other USB drives. And then an interesting concept in the upper-right hand side of that particular slide is a concept that they described as allowing the transport of content to or from TVs using a version of the USB interface for consumer products.

Moving on to slide 12, we're showing a slew of optical disks, in this case Maxell, at the Maxell booth. There are many different formats and form factors available. On slide 13 we're showing the blue laser optical disk format wars, in this case we have external devices hooked up to the same type of HP computer. In the case to the left it shows a Blu-ray optical disk hooked up here, and on the case on the right is an external HD-DVD player. Of course internal versions of these products are available. Expect over the next couple of years that the price of these drives will be decreasing enormously as the volume goes up for both formats.

Slide 14. Optical disk technology has shown future developments that could increase capacity. One of the more interesting products that have shown up at the very end of 2006 and beginning of 2007, is the introduction of holographic storage devices on a commercial basis. Maxell is producing a product that was designed by InPhase technology from Longmont, Colorado that provides a cartridge 300 gigabyte holographic write-once disk. The first use for this product is for professional video, both capture as well as archiving. About this right now, these are initial products. Hardware is still fairly bulky, and we probably will not see consumer devices available until there's additional miniaturization. But this technology does offer a path to, for instance, one terabyte consumer electronic content distribution disks in the future. I should also mention that I've also seen non-public companies that are talking about alternative technologies for advancing optical storage that could allow multiple levels of storage with capacities of a terabyte or higher as well. So it appears there's still some legs on the optical storage technology. And if content resolution is increased, then it's possible that content distribution using optical disks could still be attractive, making these products an alternative to downloading, particularly where high bandwidth would be required.

The next slide, slide 15, gives a little more information on this initial holographic storage product. On this slide it's a little bit difficult to see, but the drive is selling for about $18, 000. It's a bit pricey right now for 300 gigabyte capacity. It has a 20 megabyte per second or a 160 megabit per second read-transfer rate. It's a write-once product. The media is selling for about $180 dollars, again it's a write-once 130 millimeter disk, 3-year prior-to-recording shelf life, greater than 50-year archival life that's projected on this, without any special handling. The InPhase holographic road map talks about a product going from the 300 gigabyte, 20 megabyte per second up to an 800 gigabyte, 80 megabyte per second in the next few years, and ultimately be able to go into a 1.6 terabyte, 120 megabyte per second.

Now moving on to hard disk drive technologies that were shown, we see here a slide from Hitachi's presentation on their first one terabyte hard disk drive. This is a five-disk product. After this announcement was made, Seagate announced that by the end of the first half of 2007 they will be shipping their own one terabyte product, I believe with four disks.

That was slide 16. Moving on to Slide 17, there's some additional hard disk drive advances that perhaps have been a bit overshadowed by the one terabyte, three-and-a-half inch hard drive announcement. Fujitsu introduced at the end of 2006 a 300 gigabyte two-and-a-half inch drive and slightly out of form factor from the traditional notebook by a few millimeters, from the traditional notebook two-and-a-half inch. Toshiba introduced a 100 GB 1.8 inch drive. Seagate said they would ship one terabyte three-and-a-half inch drives the first half of 2007. Also, by the way, Seagate basically announced a 20 gigabyte one inch drive being available this year as well, as part of their announcement of the DAVE product. In mobile two-and-a-half inch hard disk drives, which are probably the single-fasted growing segment of the market for hard disk drives right now, we saw a couple of interesting developments. One is Seagate has introduced their FDE drives which do in-drive encryption. That is, the key and the encrypted data are kept in the hard disk drive, making it virtually impossible to hack or get access to it. It also allows you to, by overwriting the key, effectively eliminate access to the data on the drive. So it's a very interesting security-type product. This is currently offered by Seagate, but their product should be in line and meet specifications of the trusted computing group which will include all the other hard drive manufacturers in this space. Additionally, on two-and-a-half inch hard disk drives, with the introduction of Vista, the hybrid or ReadyDrive products with a non-volatile cache used for the hard disk drive write cache, are at least available. Whether they'll be picked up this year depends upon how much it's being promoted by Vista. At this point it sounds like this will not be a requirement until 2008 on notebook computers. That's slide 17.

Moving into slide 18, some additional developments in hard disk drive technology. Moving the SATA interface from not just being an internal drive interface for consumer and computer applications, but also using an external version of that using eSATA, data of three gigabits per second can be sustained for external storage devices which is much higher than USB and considerably higher even than the fastest FireWire or IEEE 1394 products out now. One example of a company involved in this is Silicon Image. They're talking about using the port multiplier capability to create virtualization of say a couple of hard disk drives off a single eSATA port on the motherboard. By the way, currently the products available with eSATA are primarily host bus in after-products, but native implementation of this capability in the motherboards is happening at many companies, both on computers and also on consumer products.

This port multiplication also allows you to effectively daisy-chain hard disk drives off a single eSATA port, as shown on slide 19. This can be used in the consumer application to create the capability of attaching multiple hard disk drives as an external storage device for the expansion of digital video recorders. In effect this creates the capability of instead of having to erase content to make room on your digital video recorder, if you wish to retain it you can instead plug in one of these external devices and move that older content off to that external device and play it back when you need it.

In fact, eSATA connectivity is available on many of the set-top boxes from companies such as Scientific Atlanta and Pace, and these are in fact mostly active if they are allowed by the service provider. This could provide a way of lowering the cost of set-top boxes with digital video recorder capability but allowing, for instance, an aftermarket of buying storage devices and then connecting them up. So the initial capitalization cost of the set-top boxes to the MSO or the service provider could be lowered, but allowing expansion at the discretion and expense of consumers to buy external storage to allow them to retain content.

Slide 20. Other things about this particular eSATA implementation with Silicon Image, it also allows this auto drive locking, which allows you to lock a storage device into its connected system. This could be used as a DRM technology. I have mixed feelings about this particular approach. It can be used to provide security for content or to prevent theft of content or as digital rights management, but there are potential issues if you upgrade from one device to another. If you've got the device locked into the prior device, how do you move that lock? That's slide 20.

Looking at what's happening and what's been happening in the historical perspective on the growth in use of devices with storage in the home and their connectivity, let's look back ten years to 1997. Moving on to slide 21 we show and image of, say, what a typical home might look like in 1997. It has dial-up internet connection to a home computer which maybe has a direct connection to a peripheral device such as a printer. This is the extent basically of the home network.

Perhaps you would have another computer with an Ethernet connection between that would allow another computer to access the internet as well, and that typically would be the extent of the real networking. In addition, you might have an analog videocassette recorder hooked up to a TV. Maybe you could say that you've got two different elements of connected devices, at least in the home; one being the computer and the other this analog entertainment network.

By 2007 a number of things had changed and we had some additional connectivity options. Instead of a dial-up modem, direct Ethernet connections through a cable or DSL line to a router which may have wireless capability in the home were common. Laptops for instance might be able to access wireless and be able to move into the home to this Ethernet connection, so it's used as a means of sharing the Ethernet connection. A desktop computer might be hardwired to that router.

We also, backing up, especially with direct attached external storage devices becoming more common. And this might constitute then, basically the capability of a home network. Occasionally you might have direct connections of audio/video players to a desktop computer in order to store information in the cache, which then can be played out.

Separately, cell phones have become more common. And these are becoming, although still in its infancy, allowing multimedia capabilities, playing video or audio on them, other enhanced or streaming, other enhanced features. Usually though, not connected in with this computer network.

Another separate network in the home is a digital video recorder or a set-top box that's connected to a cable or a satellite.

So in 2007 it's not atypical to see three separate and more or less isolated networks in the home; one being the computer network, cell phones, and then also the digital video recorders/set-top box. That's slide 22.

Moving on to slide 23. So this is where things stand for a typical home today. What's coming next? Well, there are reports of increased network activity worldwide. More and more people getting online and actually sharing content and that sort of thing. Therefore, severe content creation and sharing expected. And that, we believe, will bring strong storage growth trends.

Moving on to slide 24, here are some projections on the growth of home network storage. My projections for external home and small business storage out to 2012 is growing, as well as home NAS--that is storage devices that's connected to your network that allow file level access. By 2012 close to 80% of external home and small business storage could be NAS, however, only about 15% of homes and network connections are estimated to have NAS by 2012 unless some changes are made.

Looking at slide 25, this is presented at Storage Visions conference by folks from HP talking about worldwide home network installations out to 2010 but showing that home NAS percentage may be only about 7.2% by that time, increasing from about 2.2% in 2006. So there's much faster growth of the home network installations, and very small penetration of the shared storage.

Again, what could be done to increase this percentage? Well, let's take a look at things that would increase the value of network storage in the home and therefore could increase the penetration. By 2017 I could see one overall. Where in 2007 we had three generally isolated networks, in 2017 we can create, if you will, a capability of creating an overall overarching network that allows virtualization effectively of all the storage device in and around the home, handling them as a single unit.

Then also, as we see here, providing back a connection to the internet of sharing that content outside the home as well as creating capability of internet backup of that personal content to create an effective disaster recovery solution.

So, direct connect internet connections in the home with increasing capacities into routing capability with wireless networking. Allowing, for instance, laptops or whatever a laptop has become by 2017; allowing automobiles with their own wireless networking capability to tie in for entertainment and/or navigation aids. We could have multimedia cell phones and A/V players and that sort of thing that will be wirelessly able to connect up to this home network.

Maybe we have even something like Seagate's DAVE or the LSI BluONYX product that provide what I'm calling personal area network storage device, that can provide access to higher levels of content even while you're on the road, to your cell phone or even an AP player.

We have desktop computers that can be, say, wired connect up to that router. And increasingly, external storage devices rather than being direct connect could actually be connected into the network so they can be shared. Also connection, wireless or wired, of the set-top box and the cable or satellite into this overall computer network.

So, taking these islands of connectivity and of storage, turning them into a single virtualized island I think is the trend. But since content and storage are shared, this may allow some aggregation or make sense of storage de-duplication.

But it also, because it is relatively inexpensive and easy to come by storage, we will keep many copies in many places as is appropriate, as long as they can be managed and organized, backed up, etc. Storage is cheap and capacities are largely making that possible, but the content needs to be managed, indexed, and automatically backed up as part of the system.

So, trying to fit, to create a model to describe the impact of networking and sharing on the growth in storage went back to some of the models for networking. Metcalfe's law is a law that was developed in the early days of networking to describe what's called the value of a telecommunications network. The basic statement was that value is proportional to the square of the number of users of the system, that is, how many people are there and connected.

There are those who say that the actual value, which is a rather ambiguous word, is less than that. But there are also those who say that, as in slide 28, Reed's law, that the value of networking is even larger than Metcalfe's law, because not only is a member connected to the entire network, but also the network is composed of subsets of the network which interact with each other, which add value independent of either the individual or the network.

Rather than talking about, say, an ill-defined concept like value, I thought perhaps there is some way of looking at this value effectively translating into the digital storage requirements. So the next slide, 29, is an initial attempt at describing how moving from a home with many applications that are shared by individuals in the home, such as video, photos, documents, music, and games, we can create, using the internet, create connections between multiple homes to create small communities.

For instance, family groups, friends, or even temporary associations of people playing games connected to the web and also connected to others. Multiple homes could be connected here, multiple copies, multiple connection paths, multiple subnets, similar to Reed's law. There is therefore a greater potential for interaction exchange and copies of content in multiple places, which might be temporary copies or working copies, and eventually even archive copies if the perceived value is there.

Due to the creation of copies and the encouragement of new content creation sharing, my estimate is that storage growth could increase by something like an equation where S is the amount of storage created by this sharing--potential storage--equals the average size of the shared content multiplied by N-1 where N is the number of connections--minus one because you don't connect to yourself--raised to the power of M where M is equal to some group participation factor, that is, interaction times the number of people interacting.

So, effectively what we've got here is an increase in storage requirements that grow N to the Nth power. Not N squared, but N to the Nth power--potential enormous growth in storage demand that could come into being because of greater interaction networking.

Moving on to slide 30. So, what are some of these ways in which things that are driving the sharing of information that could lead to the potential of such large storage capacity increases. Which, by the way, would be of many different formats, all the way from flash to hard disk drives to even optical storage.

Social networking is an important driver of this. Examples of this: Second Life, where there are even real economic models with real-world returns even though it's done all in cyberspace. Content sharing technologies, YouTube and similar things. And then ways of creating bandwidth aggregation using networks of users, such as BitTorrent creates, as an example. Peer-to-peer bandwidth aggregation technologies making it easy to share content.

Content creation is cheap, anyone can post to the web, make any amount of content increasing enormously. We may ultimately see new modes of bringing people and ideas together to create common interests. I believe that by the next decade this could be an enormous driver to economic growth. In fact, I think that the whole personal content market and storage and other technologies supporting it will be enormous parts of the economy in the next decade.

But moving on to slide 31, how do we design storage in this new networked market that meet the needs of users? First, here are some rules that I've come up with.

First, use the most cost-effective storage components that provide enough capacity for the application.

Second, never design a product where you intentionally limit the available storage capacity to the customer. Always allow a means of a storage capacity expansion.

This is the basic issue I have with purely solid-state NAND replacement of hard disk drives, in that I don't think it gives you that capacity expansion. I think there are ways, which I'll get into shortly, that can describe how that can be done, but by itself this all or nothing approach, I think, will limit the amount of volume penetration.

Make it easy to backup and copy data. Storage is cheap; time is not. Again, this is getting into that virtualization and management of all the data in the home.

Make management and organization automatic. For instance, protect data and prevent replication of corrupt data.

Design to provide the lowest total product cost. Which, I think, ultimately could lead to storage integration concepts.

So, some interesting new concepts in mass storage, moving on to 32. One example was at the G3 conference in Barcelona. Seagate introduced their DAVE technology, which is an extension of their Free Agent products, where they announced what I'm calling a personal area network mobile storage device, allowing wireless connectivity on its own battery power, of a mass-storage device that, in the initial announced cases, of 20 Gigabyte one-inch drive that can be accessed by a nearby mobile device such as a cell phone.

I should note that LSI, or Agere, at the Storage Visions conference discussed a product called BluONYX offering similar connectivity.

For the developments combining flash and hard disk drives for instance--and this is where I'm getting into ways in which you might be able to get higher penetration of solid-state drives--could the laptop be turned, effectively, inside out with an embedded flash and a small hard disk drive as a removal module, or connect to the laptop via a PAN like Seagate's DAVE or LSI Logic's BluONYX.

Slide 33 talks a bit about those two concepts. Wireless access to a local mass-storage device offering many Gigabytes of storage could use 1-inch or 1.8-inch drives. If it was a 1.8-inch drive at this point we could create that with 100 Gigabytes capacity. The device contains its own power and therefore does not draw power from the mobile device that accesses it. This concept frees up issues around the design of storage, especially large mass-storage in the mobile product itself.

34 shows a picture of the BluONYX concept, which would have Bluetooth, WiFi, SD, or USB connectivity. About the size of a credit card, it could have either a 1-inch hard disk drive in it with up to 40 Gigabytes capacity let's say, or a flash memory.

Moving on into slide 36--how could these concepts be used in laptops to increase total laptop penetration? I think that without, my own, rather than 20% market share, I think that the market share will be less than 10% for solid-state drives in laptops by 2010. But I think it could be higher.

Here's the way I think it could be made higher. First of all, again, per those rules I talked about: Don't limit the amount of local storage available to the user. Enable choices in content access rather than restrict it. Use the storage hierarchy, that is, the characteristics of different types of storage devices, either singly or together, to your advantage.

I think that the net would be an increase in market share to greater than 15% by 2011.

This idea of turning the laptop inside out for instance; instead of NAND devices being removable storage what if they are fixed storage. Instead of the hard disk drive being a fixed storage device, it becomes removable or external. It gives the users more options. And if this combined hard disk drive and flash memory approach was designed together it could be very sleek and attractive.

So the figures at the bottom on the left-hand side show an external detachable hard disk drive approach to this, where either a wired or wireless connection of the hard disk drive was available to augment the solid-state laptop capacity. Or, on the right-hand side where the hard disk drive could be a module that you could move in or out of the laptop computer as you need it. Or even a combination of these.

So, in general, moving on to 37, some changes in the electronics in the market. Customers don't buy anymore just because a product has a faster processor or because it has more storage. Instead, there are new trends driving the consumer electronic market. There is mass customization, that is, low-cost products that actually meet individual needs. There are more niche products, a loss of mainstream. More and more products and content have to access individual preferences and experiences.

Developing new levels of personal storage could drive entirely new markets for commercial products; utilize this information to the benefit of individuals. More storage at this time is shipped for personal use--we include personal computers--than for commercial, that is, IT use. And on the commercial side, more storage is generated automatically by machines rather than by people.

These are the trends; and we must design and market and analyze that market with these trends in mind.

Slide 38. The personal content will become more and more important in the future. By the next decade there will be more personal than commercial digital content stored; creating new markets and opportunities to serve this diverse and dispersed market.

Increasingly, consumer electronics are driving new technologies. Consumer electronics could displace IT in the generation of technology solutions.

Storage demand essentially is infinite; we can't keep enough information. As long as we can afford it--find it, keep it, and preserve it.

Storage must help in organizing and finding data. An example might be object storage to enhance file system accessing in consumer electronic devices.

Looking at the trends in home storage, at the 2006 Storage Visions conference we made a projection that if we look at the trends on commercial and personal content in a typical tech-savvy home we might end up with something about 4.5 terabytes total content by 2010, a little bit less than half of this being personal content. That's on slide 39.

Moving to slide 40. At the 2007 Storage Visions conference, and assuming that with the data being the same up to 2010, but assuming in 2010 that somebody in that household starts to record their life as it happens. With certain assumptions on the frequency of that sampling and the resolution of that sampling that's being done, we could easily see accumulated storage capacities of hundreds of Terabytes by the next decade; making it easy to consider that an individual may accumulate much more than a petabyte, maybe even multiple petabytes of data in their lifetime.

So, the implications, at least out to 2012, in terms of hard disk drive market niche projections are shown in slide 41 where we show mobile consumer electronic, desktop enterprise ATA, and traditional Fibre Channel SCSI enterprise. We see starting in 2010, mobile computers exceeding desktop, and in fact are the fastest growing market.

By 2013 I project that over one billion disk drives will be shipped per year. 2006 it was about 436 million units, 2007 should be approaching about 500 million units.

Moving on to 42, looking at the hard disk drive form-factors in 2012. After 2010 2.5-inch and smaller hard disk drives, I believe, will dominate the 3.5-inch. And higher capacity and smaller form-factor needed by mobile storage will be driving this trend.

Ultimately, moving to slide 43, by 2015 I think we will see a terabyte in your pocket, a petabyte in your home, exabytes in data-centers, zetabytes in the world. Each of these is a factor of 1, 000 times the other. Because digital storage demand is very elastic. So that's the end of 43.

I'd like to thank you for listening. I enjoyed talking with you and getting a chance to meet with you. If you have any questions, I'd be glad to discuss with you further. Take care. Tom Coughlin. Bye.

---