The Case for Storage Post rated (max 5)
By Christopher Thiessen (rating 3.59) on 12:21 02.09.01
With shares of most leading storage firms holding up relatively well through the recent market doldrums, its time to see what all of the excitement is about. I will start off with a discussion of some of the estimates for the storage market, and then I'll consider the sector's driving factors.
Some Facts
These days, there is a lot of attention being focused on the storage market by investors; it's pretty easy to see why. Research firm Dataquest estimates that the overall storage market will reach $72 billion in the year 2003, up from the $32 billion spent in 1999, which represents an annual growth rate of 24.7%. Dataquest also estimates that 1.3 million terabytes of storage will be purchased in the year 2003, up from the 250,000 terabytes purchased last year. It could get even more exciting. Based on internal estimates, EMC thinks its possible that storage capacity estimates are as much as 20,000 times too low for 2010. Chew on that for a while...
Drivers
Let's look at some of the drivers for the storage market. We have had explosive growth in a number of applications. The Internet, data warehousing, e-mail, digital pictures, mp3's, MPEG's all require vast amounts of storage. These technologies are all in very early stages of their use, and will grow exponentially for a number of years to come. Future applications, such as voice/speech recognition, streaming video, video on demand, video conferencing, web TV will also be key drivers in storage demand. And this is just the tip of the proverbial iceberg.
The CEO of EMC spoke at the Merrill Lynch Storage Conference at the end of January, and presented a very compelling slide which pretty much sums up the opportunity in storage. Technology spending and development has tended to move in cycles, with each successive cycle creating 10x more value than the previous cycle. We are currently in what is being called the networking cycle, and companies like Cisco, Juniper, Nortel, Corning, and JDS Uniphase are the main beneficiaries of the build-out of networks. In 2010, the information, or content, cycle will take over as the prominant creator of value. Companies will spring up to fill the pipes of the networks currently being put in place, taking advantage of the vast amounts of bandwidth. As this content is developed, digitally, it must be stored somewhere, and that is why storage jumps out like a Jack In the Box.
Storage Networking
So far, we have only been looking at pure storage, i.e. hard disks, zip drives, cd's, etc. One of the most exciting areas of the storage sector is storage networking. Why? The economics of storage networks are hard to ignore. They improve the efficiency of corporate local area networks (LAN's). They can offload resources from the LAN to a storage network. And they cut the amount of manpower needed to administer storage.
Storage networks come in two basic flavors: Storage Area Networks (SAN) and Network Attached Storage (NAS). Confusing? When people start talking in acronyms, this becomes almost impossible to follow, and you will soon see that it is only the beginning of the confusion.
Let's start with something a little more conventional, the old school version of how storage was integrated into a corporate LAN environment, Direct Attached Storage (DAS). Historically, corporations would have file systems attached directly to servers, creating a direct connection to the LAN. This was okay because it was simple, just hook the file and tape backup systems directly to the servers, and voila, you're set. When a user wants to access information that is on the network, it has to be accessed through the specific server that is attached to the file system on which the information is stored. When data is backed up on the network, it must pass through two servers, the one attached to the tape backup system, and the one attached to the file system. This takes up LAN resources.
NAS
Network Attached Storage is similar to DAS in that it connects directly to a LAN. The main difference is NAS implements a specialized NAS Appliance which takes over the duties of the servers, freeing up LAN resources. The file-serving duties are processed by the specialized NAS appliance. The NAS appliances typically have file systems attached directly to them, and are highly scalable. Since NAS appliances are specialized specifically for storage, they are not required to do other multi-purpose tasks, such as print serving, or web-serving, which increases the performance of the storage network. NAS is usually used in lower end storage networks, as so far it has been hard to scale in a cost-effective manner.
SAN
Storage Area Networks are, simply, separate networks set up to handle the storage needs of a corporation. The most simple way to envision a SAN is to think of a switch lying between each server and each file system, with all switches interconnected to each other. (Or, you can just look at the graphic below.) This allows for what is known as any-to-any connectivity, with any server able to access information stored on any file system. The major benefits of a SAN system are: 1) any server can access information stored on any disk array (if one server goes down, data on all disk arrays is still available), 2) backups are done without utilizing a server or the LAN, freeing up LAN resources for other purposes, 3) SAN's are easily scalable.
NAS vs. SAN, or the Convergence of the Two
There has been a lot of controversy as to which storage networking system was better. NAS proponents say NAS is cheaper, easier to administer and maintain. Proponents of SAN say its more scalable, has higher transfer speeds, and removes all congestion from the LAN.
But, if we think about it a bit, isn't it logical that the two merge into one? Essentially, when NAS environments scale, they become SAN's. And, why not combine the two, starting with a NAS appliance, building out a SAN behind it to provide redundant storage capabilities?
We are already beginning to see the combination of the two architectures. Network Appliance has recently begun incorporating Brocade Silkworm technology into its NAS appliances (Brocade is a provider of SAN switches, based on Fibre Channel). EMC's high end NAS offering is essentially a NAS with SAN architecture. By the end of this year, I would be surprised if there was any distinction made between the two.
Summary
The storage market will be an exciting place to be for the next decade. All signs point to a long term boom in the storage sector, creating much more value than what we have seen thus far. And the trends tell me that the valuations currently being put on the companies in the storage sector are, for the most part, justified. If you are a long term investor, these short term over-valuations and short term economic problems should not even appear as more than a blip on the screen.
However, I think that we are seeing the signs that, despite being deemed a "must spend" technology by corporate IT managers, the short term economic situations/fluctuations will have some impact on storage spending. Qlogic stated earlier this week that they were beginning to see a slight inventory build up, and slight increases in DSO's. Network Appliance just reported revenue numbers that were slightly below estimates. They said that everything was typical going into the holiday season, but the immediate pickup usually seen after the holiday season did not happen. In fact, the CEO of NTAP said the first 10 days after the holidays were dead. But things picked up significantly after that, and they saw demand come back to more normal levels. Nevertheless, NTAP told analysts to expect revenue growth at the low end of the estimates they previously set. With this as a backdrop, the near term could see some pressure for the names in the storage sector. The question then becomes, how much of this potential slowdown is priced into these stocks?
I fully expect a number of winners to come out of the current environment. NTAP looks like a solid bet to be a good investment for those with more than a 6 month time frame. EMC also looks like a winner. Brocade looks strong, as does Veritas on the software side. I think McDATA has a chance to be something, but they will have to make some strong inroads into Brocade's business to do so. I also think that outsourced storage vendors (such as StorageNetworks, Storage Way, and Storability) as a group will make a strong showing, but I'm not yet convinced that any particular outsourcing company is going to be a true winner.
By Christopher Thiessen MetaMarkets.com
community.metamarkets.com
Link has graphics.
Jack |
