Blog Host: Randy Kerns - Independent storage consultant READ ENTIRE BIO

The tough part is figuring out what the value is to your company (and to you getting your job done, on more of a personal level) for those features. If you're not going to use them, then you should probably disregard them. Usually they aren't really free. There's some tradeoff for price or complexity or weakness in another area.

We often ask customers we work with to apply a weighting factor to several different areas of a product. Doing a comparison to figure out which product meets the need becomes a little easier then because you can evaluate and score a product in particular areas and then apply the customer weighting factor to see if it's really meaningful and come up with a composite value. Really, the weighting factor is common sense: How much do you value this particular feature or characteristic? It's worth the effort.
Posted by Randy Kerns Product differentiation
13 JAN 2005 05:36 EST (10:36, GMT)
When comparing product offerings from different vendors, there are usually some differences beyond price. Features, performance, connectivity and so on are all things that we look at. Depending on the type of solution, there may be more differentiating elements. What that means is: Comparing a NAS device to a block storage SAN device would have fundamental differences that make device comparison more of a secondary issue. It is more important to look at a solution level and see how one particular method would meet the requirements vs. another. It becomes even more complex with an object-based storage device because of the abstraction of storage location and identification that today is exploited by specific applications.

It comes down to a top-down approach at looking at solutions to meet requirements. First the requirements have to be understood, and then the different approaches and the infrastructures involved need to be analyzed for their ability to meet the requirements. Once that is resolved, then individual vendor offerings and the characteristics can be examined. From that, a comparison can be done and the differences can be examined to help guide a customer in making the right purchasing decision. There are some free comparison workbooks available on our Web site that might help. Ultimately, it boils down to this: You have to do your homework or engage someone to help you do it.
Posted by Randy Kerns Abused words
12 JAN 2005 06:25 EST (11:25, GMT)
I read a great deal of product material from vendors, which includes technical information as well as marketing collaterals. I also get marketing presentations. It's funny, but there are different phrases or words that become used by more than one company as if they are clothes coming into fashion and then becoming passé. Some of the words get overused to the point of becoming almost comical. It's sort of like the joke about the comedians that get together and just say the joke number and they all laugh -- they know the punchline so well.

Over-using the marketing phrases or words probably works against a marketing message. It leaves an image that what is being said is not fresh or new. Sort of an old news message. Why is it that vendors feel compelled to use phrases that are simply worn out? Most of the phrases are really analogies that were coined to make an association to aid in understanding.

Let's take one phrase that has been beaten to death: forklift upgrade. That phrase does conjure up an image. But I don't think there are many data centers that exist (if any) where you can really use a forklift. So it's an analogy of moving something big. How relevant is that today? Maybe with storage systems I really do want to replace them after the warranty period is up because of technology changes. The new models have really made the old ones obsolete. So the phrase forklift upgrade, which is used in a pejorative manner, is really not communicating what the real situation is for a customer. But vendors seem to be stuck on using it like the phrase shows that their product has a longevity others may not. That may not be the case.

I like it better with either a fresh description of product characteristics and value or at least accuracy.
Posted by Randy Kerns Credibility
11 JAN 2005 05:27 EST (10:27, GMT)
We always talk about making wise choices regarding storage or storage management software purchases. It is usually a difficult decision because there are so many choices and so many factors that are not simple A vs. B comparisons. We help a lot of people evaluate the products, company strategies and technology and hopefully give them helpful advice. One of the problems I often run into is that someone has "heard" or gotten information that is not quite right. It can be "not quite right" enough to really mislead them or send them down a path that is just wrong for what they need.

Undoing that impression or stilted information is much harder than explaining why things are the way they are and what the attributes are of specific solutions. Sometimes I end up having to talk about the credibility of the source of the information. That is pretty easy, actually. It's just a matter of talking about the experience someone might have had. Was it really in storage? Was it in a position that is meaningful to you? How much experience? What is their background? Do they have some vested interested in a vendor? Tough questions sometimes, but they need to be asked when there has been misinformation given.

The easiest advice is always, "Vet your sources." Do you believe everything that someone might have written in a trade magazine? Lots of simple things to ask. I'm a skeptical person anyway, so it's easy to pass along the questions I would ask. Most of the decisions we make with storage are crucial, and we have to live with them for a long time. It is best to be careful.
Posted by Randy Kerns Storage performance
10 JAN 2005 20:58 EST (01:58, GMT)
I get involved in the discussions about performance for storage systems quite often. For block storage systems (those that do block I/O, which is usually direct-attached or SAN-attached storage), I always recommend looking at the Storage Performance Council performance results, which contains credible results of a performance metric that represents a real-world environment and is audited by an independent party to ensure that the results are valid and the configuration is representative. In addition, the dollars per I/O is calculated.

Not all vendors have their performance evaluated with the SPC benchmark. That is a shame, and we recommend that customers require that information. Many vendors try to give their own benchmark data or pay some organization (usually a testing lab) to run the tests that they specify in a specific configuration. I'm always very skeptical of these types of results -- I want an independent benchmark that has been audited.

The SPC-1 benchmark represents an online transaction processing environment like you would typically see with some databases. Additional benchmarks are in some stage of development. Again, I highly recommend customers look to this data for block storage system performance information. There is extensive information about the benchmark on the Web site and customers can buy their own copy of the benchmark to run if they want.
Posted by Randy Kerns Storage purchases
07 JAN 2005 11:22 EST (16:22, GMT)
There are trends on how storage is purchased and a few of them are worth pointing out. One is that in the enterprise data center market segment, storage is purchased independently of the server for the most part. In that market, the storage purchase is centralized and done on a fairly regular basis -- usually once or twice a year. Usually a customer will purchase from only a small set of vendors at any one time -- two in most cases and only changes the vendor selection about every three years. It's pretty tough to break into that cycle.

In the small to medium-sized business market, the primary storage is almost always purchased with the server purchase, which is termed a "server-led sale." There is a secondary storage purchase that occurs and for the most part, the customer will purchase from the same vendor. But not always. There is great opportunity here for a vendor to make a sale with that secondary storage purchase. And it's a big market that is growing faster than any other for storage.

It is important to evaluate the storage products AND the storage vendor before making the selection. Buying storage at the same time as the server may not be the best move -- there may be a better storage solution -- but a package deal may make it compelling. I always recommend evaluating storage independently and not getting locked into a storage solution that might be a compromise.
Posted by Randy Kerns Negative marketing
06 JAN 2005 11:45 EST (16:45, GMT)
It is always best to look at the attributes of a product and determine whether it meets the requirements. What you want to see is how a particular product can solve your problem and what its merits are. In addition, a comparison between competing products is useful to help make the best choice.

Unfortunately, some of the material that is used to represent a product focuses on the negative aspects of competing products and then why how the product they are proposing doesn't have those negative characteristics. What a bad way to represent a product! Rather than starting off with the negative, getting the customer to think negatively and then proposing an answer to the negative things, the material should start with the positive attributes and then have a comparison.

Starting with the negative either means that there really isn't a significant advantage for the product being proposed or somebody just doesn't understand how to market a product to intelligent people. I find that people who deal with storage are some of the brightest (but overworked) people. They quickly see through a negative marketing approach. So why do we keep seeing it? I guess it's just the easiest way to represent a product. A little more thought, more effort and more knowledge, though, might lead to a better method of representing a product.
Posted by Randy Kerns Upgrading storage systems
05 JAN 2005 20:05 EST (01:05, GMT)
Vendors sometimes have marketing messages about being able to upgrade from one storage system to the new model. A new model storage system usually means a new controller that is faster with more cache and maybe more interface connections. The upgrade may vary from just replacing the controllers (or cards within the controller) to moving the disk drives to an entirely new box. This degree of variability can result in widely different customer experiences.

Being able to upgrade the storage represents a significant cost to the vendor as well. The vendor has to design and develop the upgrade capability which may be hardware as well as software. There is a non-trivial testing period required for all the different combinations of upgrades as well as the training for support and field people. Field stocking of parts for upgrade is a major cost as well.

So why have the upgrade as an capability with a new product offering? Is it such a valuable feature that vendors must have it? Do customers really do it? The truth is that an upgrade of storage systems happens rarely. Storage systems may have a 3- or 4- or 5-year useful lifespan due to technology changes, so customers usually buy new (additional) storage and retire the older systems when it makes economic sense. The storage world has changed, so upgrading storage systems by and large is an archaic concept.

When the marketing guy says that the storage system can be upgraded, there are some significant costs to enable that capability that you have to pay for. And, in reality, you're probably not going to do it. So don't ask for it or make a decision based upon that unless you really have a need for it.
Posted by Randy Kerns Confusion on storage networking
04 JAN 2005 05:46 EST (10:46, GMT)
The questions I received for today were all about networking, which really isn't an area that I feel comfortable (or have a desire for) answering. The problem is related to word order and association. Storage networking means putting networking storage systems together -- usually with a fibre channel infrastructure, but there have been some deployments over IP with iSCSI. Unfortunately, some people (and some companies) call it "network storage," which keys a lot of people to think of networking and ignore the storage aspect. I think it would be better, just because of the way people latch onto the first word, if we used storage network instead.

There are many reasons for that. Typically a network guy doesn't really understand storage. They don't understand LUN allocation, affinity groups/allegiance, LUN masking, load balancing drivers and other very arcane things. And, conversely, storage guys don't usually mess with networks. So if you jump to the conclusion that network storage is the networking guy, that probably won't get you where you want to go.

Which leads to another issue. In enterprises that are of some size, there are always storage professionals. In the larger ones, there is usually a department that handles storage networks. Typically these are professionals with some special training or experience. In smaller environments, there may be one person (or a part-time person) to deal with storage. But they usually have some hard-earned experience that is invaluable.
Posted by Randy Kerns Demand for capacity
03 JAN 2005 15:22 EST (20:22, GMT)
The storage business continues to be a very interesting area. There are new products and techniques, but the problems are still basically the same. The biggest problem that customers have to deal with today is the demand for capacity. It results in more storage being deployed, but there's not enough additional budget to hire storage people to administer all that new capacity. Vendors are addressing this problem with many different solutions. Some are specific devices, while others are software for management and still more are combinations.

What a customer ultimately buys depends on many different factors. The biggest issue for a customer is to always understand the requirements. From that, a logical evaluation can be done to arrive at a small set of solutions to choose from.

During my stint on answering questions here, feel free to ask about storage issues and strategies. It will be difficult to answer some questions where the environment or requirements aren't understood, but I will try.
Posted by Randy Kerns