Skip to main content
Message Font: Serif | Sans-Serif
No. of Recommendations: 48
Thanks, xu4054, for the interesting thread on Procom on the NTAP board:

It's worth a read. The NTAP board is a good place to learn about storage systems in general.

I've also wanted to make a purchase in storage. I've decided to make a tiny purchase today. I'll summarize my story, with a Gilder slant, here:


Back in November, Gilder attached his name to a new new paradigm: "Storewidth". Although he hasn't added any networked storage companies to his list, he has discussed the business.

On storage in general he revaled the foundations of his thinking with two tech predictions, damning the technologies on either side of hard disks:

(1) silicon memory will not replace disks, because a tiny magnetic domain on a disk is inherently simpler to manufacture than a silicon capacitor and its supporting wires on a chip.

(2) a tape hierarchy will not replace disks, because disks are getting hugely abundant and cheap - and they are fast.

(to quote Gilder) "The rapidly collapsing price of storage dictates architectures that waste storage and economize on processing and customer time. Intelligent hierarchical storage systems did just the opposite. As in bandwidth so in storewidth, abundance trumps intelligence every time."

So disks are important. This explains the rise of EMC, right? But it doesn't look like he's adding EMC to his favorites.


In November, Gilder also explained the division between Network Attached Storage (NAS) sold by companies like NTAP, where the disk is attached to a very simple, old-fashioned Ethernet technology, and Storage Area Networks (SAN) sold by companies like EMC, where the disk is attached to a new-fangled, sophisticated Fibre Channel Network.

This division has been getting buzz lately. So it's worth understanding what's behind the technology.

You can read about Fibre Channel here - it was sort of a network & SCSI replacement in one:

Fibre Channel is one of these Guaranteed Delivery protocols that Gilder loves to pick on. Fibre Channel, he asserts, is too much of a headache for administrators, and except for those who need the very highest bandwidth, he predicts customers will choose NAS over SAN just to avoid Fibre Channel. Sort of reminds me of his position in the CDMA vs TDMA debate, or of the way he disparages SONET, or Quality-of-Service protocols. If collision is a problem, he asserts, the answer is more capacity, not more intricate management. Gigabit Ethernet is the future.

The benefits and liabilities of the approach taken by NAS is well explained on the NTAP board - here's excellent post, with a list of the same NAS competitors that Gilder has mentioned

Note that the SAN leader EMC is now also entering NAS.

OK, enough background. Time for stockpicking. Here are some of the NAS companies.


xu4054's thread about Procom was informative and the verdict was decidedly negative, on the note that revenue continues to fall in a growing market.

If you go to the Procom website and try to find out "what are the comparative advantages of Procom products?" the answer is largely "they're reliable"; they also have an interesting modification-rollback feature in their software. But they provide no comparative benchmarks or other numbers, and with all the competitors producing many-nines reliability, it's not clear that this kind of focus on reliability is going to win them any more share in the future.


They are the current NAS leaders. Honestly, I haven't determined yet their why product so excellent compared to the other NAS vendors. (Any information?) But they're the revenue leaders, so their product must be pretty good. Gilder noted in November that MP3.COM chose to use them for storage.

They sell servers as small as 50GB, and up to more than a terabyte.

Their revenues were $125 million last quarter, up 90% year-over-year.


the original NAS company. It has the number two position, with $22 million in revenue last quarter. This is actually declining sequentially from $26 million in the previous quarter. But they are rasing more money (via equity) to focus more energy on the NAS business. Note that they have less than 2% of the market cap of NTAP, so they may be quite a bargain if they succeed in their efforts.

Like Procom, they sell servers larger than 50GB, and they go up to 9TB per system. Their product literature claims that they are the leaders in performance when it comes to NT-compatible filesystems. Perhaps the reason they trail NTAP is that NT performance is not as important for customers as Unix filesystems. (Just a guess.)


Quantum acquied Meridian Data, and they sell low-end NAS, starting as low as $500, and no bigger than 40GB. Note that in 1999 Quantum was split into two stocks with an eye on entering the NAS business: HDD tracks the group that sells their hard drive technology, while DSS is the company that's in the NAS business (they also sell tape systems and tape media).

The $500 NAS box they sell is just 10GB, and it's their newest NAS product. They are clearly driving to dominate the low end, and the key benefit their product literature touts is low cost and ease of administration ("up and running in 15 minutes!"); performance and reliability are also important, but secondary.

The part of their revenue attributable to storage systems (rather than tape media) was $82 million last quarter, growing 50% year over year.


This company went public in November. They specialize on easy administration network appliances - their big product is the "Cobalt Qube" (a linux-based web server/file server as small as 4.3 GB). They also sell a NAS product called the NASRAQ (as small as 20GB). price on the 4.3GB Qube: $781. Price on the 20GB NAS box: $1,500. More expensive than Quantum's products per byte of storage.

As a Linux box, the Qube contains a lot of functionality that you won't find in Quantum's product: mailserver, firewall, standard Apache server, Java support, etc. Their key advantage is also ease of use rather than performance. The Fool boards report that they are focused more on the software side of the business than the hardware side.

They are small and growing fast. They reported revenues of just $9 million last quarter. This is a 45% sequential increase from last quarter, and a 350% year-to-year increase.

After doing an IPO in the $22 range, they're trading now at near $100 (down from much higher - but jumped $15 today!), which gives them a $2.5b marketcap, which makes them worth more than Quantum DSS, on a fraction of the sales. Maybe it's the growth; maybe it's the Linux hype.

I think the Qube product concept might pay off if they can find a way of locking in their customers as they move to higher-end business. With Linux, this is a serious question: barriers to entry for high-compatibility competitors seem low.

And their concept is straying from the Gilder's core belief in the "dumb storage" concept: the Qube, while simple, is pretty smart as a Linux box, packed with silicon and thousands of lines of software. And the company is pricey - you pay $62 for one dollar of trailing sales, about twice what you pay for a dollar of trailing sales at high-end leader NTAP.

Still, it might be worth a look.


In September, Maxtor acquired Creative Design Solutions for $57 million. CDS made a low-end NAS product called "Plug & Stor".

Know nothing else about this except that Quantum benchmarks against both them and Cobalt.


There are other NAS sellers at the low-to-midrange end - I haven't examined them all, but I think I've hit the most significant competitors. Some of them specialize in non-magnetic media (for example, AMS Storage specializes in CD and DVD storage, but they also have a hard-drive NAS product). Here is a buyer's guide list:


The NAS market is split into high-end and low-end businesses. NTAP is the market share leader at the high end and the one that has the most sky-high valuation right now (trailing P/E = 330, with 90% revenue growth year-over-year). The gathering buzz about "the next EMC" is helping to justify the high valuation.

Procom looks weak at the high end, and ASPX, while fairly strong and a better deal by price/sales than NTAP, is also suffering losses while in direct competition with NTAP and EMC. So their future is far from certain.

And in the long run, Christensen's rule of thumb is that it's easier to move upmarket than to move downmarket. This makes DSS and Cobalt attractive. The biggest problem facing server administrators is the human costs of administration. By focusing on making an "incredibly dumb" cheap storage box, the low-end makers may someday be able to move upmarket and win big customers.

Both DSS and Cobalt are interesting. It's wise to look at upstart Cobalt as a software company that happens to sell boxes. With the main attraction being standards that come from open source Linux, it's difficult to understand if they will have any technology lock-in by providing cheap and easy Apache HTTP and Sendmail-based mail. There is a danger that everything that's valuable in what they sell is freely copyable. A very interesting product, but, especially with the current prices, a hard call as a company and a stock.

On the other hand, to me DSS looks like it might be a pretty good deal (P/E = 17, trading near 52-week lows, with 50% growth in its NAS business). Its market cap, $1.7b makes it 1/10th the price of NTAP (at $16b) and 70% the market cap of upstart COBT, it already has reasonably solid prospects. The DSS group is profitable (helped by the tape sales no doubt), and it is seeing healthy revenue growth in the NAS business. I also like the fact that Quantum top managment is mature, and has weathered many technology shifts over the years.

And with its hardware focus, DSS has a price/performance advantage at the low end that will probably be difficult to unseat.

Whether the COBT or DSS has the better product concept is a toss up now - they're different. I just wish their products were more clearly differentiated from each other or more complimentary - then I would buy both! Right now, if you only want to buy one server box to do everything, then you would want COBT's Qube. But if you have a network that just needs a bit of extra storage, then DSS is what you would buy. COBT sells the software with cheap and easy smarts, and DSS sells the hardware with idiotic storage muscle. Howver, unless there is explicit cooperation between the two approaches, it seems likely that the companies will work to encroach on each other's businesses by improving their products to make them more similar.

I'm going to buy a little DSS today to hold on for the long run. A relatively cheap, hopeful, gem. COBT is too rich for me today.

I'd be interested if anybody bought some COBT (or NTAP, or DSS, or others), and what you think.
Print the post  


What was Your Dumbest Investment?
Share it with us -- and learn from others' stories of flubs.
When Life Gives You Lemons
We all have had hardships and made poor decisions. The important thing is how we respond and grow. Read the story of a Fool who started from nothing, and looks to gain everything.
Contact Us
Contact Customer Service and other Fool departments here.
Work for Fools?
Winner of the Washingtonian great places to work, and Glassdoor #1 Company to Work For 2015! Have access to all of TMF's online and email products for FREE, and be paid for your contributions to TMF! Click the link and start your Fool career.