Super DLT, LTO face off

Posted on June 01, 2001

RssImageAltText

Point/Counterpoint

BY HEIDI BIGGAR

No two tape technologies in recent history have generated such heated debate as have Linear-Tape Open (LTO) Ultrium and Super DLTtape. With both products now shipping in volume, analysts have begun to share their forecasts for these technologies. Once considered the underdog in this battle, LTO Ultrium is proving to be a competitive force in key markets (see "Tape library market stays strong," InfoStor, May 2001, p. 12).

Following is an excerpt from separate interviews with Barbara Nelson, president of Quantum's Super DLTtape group, and Brenda Zawatski, vice president of removable storage solutions at IBM. The discussions centered on projections made by Bob Abraham in the recently published Tape Library Outlook 2001. According to his forecasts, LTO will erode DLT's stronghold-in terms of unit shipments and revenue-by 2006 (see figure).

Click here to enlarge image

In addition to IBM, Hewlett-Packard and Seagate also manufacture LTO Ultrium drives. IBM, however, was the first to ship volume quantities of these drives and, last month, the first to deliver a Fibre Channel product. Tandberg Data is a second source of Super DLT drives.

How is the technology being received in the marketplace (e.g., OEMs, channel, etc.)?


BARBARA NELSON -- Quantum
Click here to enlarge image

Nelson: Our marketing has al-ways been through our OEM partners, library partners, etc. That way you have one brand that can be pulled through. With our channel partners, brand re-cognition is up to 65% right now. So, it's proven to be a good strategy. And particularly important, Compaq continues to be exclusive on Super DLT; they have not endorsed LTO. Compaq has a significant portion of the server market, so that really helps us out.

Zawatski: We are operating in five channels today: our sales force, business partners, high-volume distributors, OEMs, and ibm.com. We're seeing success in all channels-though some more than others. We have several OEM agreements in place already, and several others are under negotiation and are expected to close in the next 60 to 90 days.

With OEMs, there is a longer qualification cycle. Compaq has been a DLT customer. But there are Compaq competitors that may choose LTO for other reasons. Incidentally, Compaq also has some of our Anaconda Ultrium libraries in their SAN [storage area network] centers.

What has been the response among library manufacturers to Super DLT and LTO, respectively?

Nelson: Library customers tend to be faster to migrate to new DLT technologies than in some of the server-attached applications. There is certainly some replacement business. It's mixed. Particularly when you have a new product, the very initial stages are typically new customers, and then you get into the transition.

Zawatski: We're shipping to almost every library manufacturer. There are at least eight manufacturers right now-including IBM-that handle LTO. ADIC is a very close partner. ATL will soon offer an LTO product, and they're owned by Quantum. We're seeing market demand for automation providers to provide choices. And I think the fact that automation providers are supporting both LTO and DLT is an indication of the acceptance that we are seeing.

Can you provide an update on drive shipments? What do preliminary numbers suggest? Are they in sync with Freeman Reports' tape library forecast?

Nelson: Last quarter, Super DLT shipments [accounted for] less than 10% [of our total unit shipments]. We expect it to be roughly 10% of our total units this quarter and roughly one-third of our total volume for the year.

[These numbers reflect] how quickly or slowly the market moves to something new. If they are happy with something, end users tend to continue to buy that product, even though there may be something better. And even though you might think the move to Super DLT, which has three times the capacity and twice the speed, is a no-brainer, it just takes some people time to make the decision to move over.

Zawatski: I think these numbers are lining up pretty well. We've shipped more than 25,000 drives to date (since our launch Sept. 1), and more than 500 Anaconda libraries. My hope is that [Freeman's] 38% LTO share in 2006 is understated and that we actually get more.

Please comment specifically on Freeman Reports' projection that LTO will have a 38% market share, versus DLT's 43%, in 2006.

Nelson: Obviously, we think it is wrong. We see LTO with a much smaller percentage in that time. One of the critical things for customers is the ability to read years' worth of data cartridges without having to go through upgrades and changes. Customers literally have years' worth of DLT cartridges and they are going to want to be able read that data.

The fact that our installed base is so large-a million-and-a-half drives and 55 million cartridges-is tremendously im-portant. Certainly, if we don't execute, you could see our share dropping. If we execute, customers will have no reason to switch. And we have 65% awareness in the channel, versus a couple of percentage points for LTO.

Zawatski: I think Freeman Reports' projection of 38% market share for LTO in 2006 is possible. In fact, we've already moved on to the second generation. I think we have forced Quantum's hand. I think they announced before they were ready and that may be part of the reason why there has been a trickle of Super DLTs out there. We want to keep the pressure on Quantum not only from a drive standpoint, but in terms of integrating the drive in automation products and servers.

What about the cost of R&D going forward? How is that expense justified?

Nelson: These products require a lot of investment to keep the technology going. With our market share and our position, the business model makes sense for us to continue that investment because we see a return on it as a result of our installed base.

If you do the same math for our competitor, it's not as compelling, and it costs them the same amount of money. And just because our competitors are a consortium, there is no difference in the amount of money that is spent. If you look back at how much it cost IBM, HP, or Seagate to come up with their version of LTO, they're all in the same ballpark, relative to what it cost us to come out with Super DLT. And you have to stay on that curve. Customers are going with you not only because they want to see this product, but because of what they want to see one, two, three, etc., years from now.


BRENDA ZAWATSKI -- IBM
Click here to enlarge image

Zawatski: We'll continue to invest heavily. I would say our investments are a little bit less than HP's and Seagate's, [for instance,] because we leverage our high-end 3590 technology. That's why we offer additional functionality. We've got a lot of intellectual property within IBM. The installed base is obviously key, but if you've got enough value proposition in your products, customers will make the switch. You've got to make it worth their while. You've got to have the additional functions and features that customers need. I think we have that with LTO.

Backward-compatibility is a key issue. How important is maintaining backward-compatibility through a tape family?

Nelson: We will keep Super DLT backward-read-compatible through the entire road map, but the question that comes up is "how far back do you go?" We assess that at every generation because you can kill yourself by having it backward-read-compatible with everything you've ever made-which people don't care about. But you definitely want to keep the advantage within a generation, typically, of what you are doing. I would expect the same philosophy that we applied to Super DLT to apply to future generations.

Zawatski: Cartridge interchange is also important, and vendors like ATL are doing mixed-media libraries. We'll announce-probably within the next 30 to 60 days-our own mixed-media Anaconda library. This means that rather than having users migrate all their data, we're going to do that in the library for them, so that they're not spending additional time and resources and having to re-write cartridges. This library will use what users have and will take them on to better technology-LTO-very easily.

What will the "Super Drive" market look like in 2004? Will there be fewer players?

Nelson: I predict that four years from now, you won't see as many players. Vendors will either drop out or merge. I think they'll look at the business part of it and say they aren't getting the return for what they are investing. Even if Freeman Reports' projections are correct, Ultrium vendors have to divide that number three ways. I think there will be some fallout in the market.

Zawatski: The vendors that execute flawlessly, have a lot of innovation, and have a road map of where they are going will be the ones that succeed. The tape market is a growing marketplace-between 10% and 11% CAGR. But there probably will be fewer players. I think some of the smaller players will fall out just because drive technology is an expensive area in which to invest. You need not only R&D dollars, but also the talent and skills from a resource perspective and the technology know-how.


Coming in INFOSTOR...

SPECIAL REPORT: RAID arrays and controllers

null

Also in the July issue:

  • Next-generation FC SANs: 2Gbps fabrics
  • Fibre Channel, IP, InfiniBand: Which, when?
  • High availability for Windows 2000


Comment and Contribute
(Maximum characters: 1200). You have
characters left.

InfoStor Article Categories:

SAN - Storage Area Network   Disk Arrays
NAS - Network Attached Storage   Storage Blogs
Storage Management   Archived Issues
Backup and Recovery   Data Storage Archives