Storage pros everywhere: don't think you're exempt from the "X as a service" trend.
Given the explosive growth in storage and rising aggregate costs, maybe you'll end up being the *first* part of the IT organization to transition to this service-oriented model.
To be fair, not everyone should care about this particular topic -- or at least, not yet.
If your environment is big and getting bigger faster (>100TB) you might care. If your environment uses storage dynamically (lots of requests with fast turnaround expected) you ought to pay attention to this discussion.
And, despite me working for a vendor (EMC), this is going to be more about how you use the tools vs. the tools themselves.
So, let's figure out together just how far along you are on this journey.
The Rationale For Storage As Service
I'd like to share a few oft-repeated thoughts before we begin.
For every gigabyte you managed in 2010, expect 50x more by the end of the decade. That is, unless you're in an information-intensive industry, and then you should expect much more.
For every server instance you're running today, expect 10x more by the end of the decade.
Unless, of course, your business makes heavy use of internal and external applications.
You're not going to be able to get a lot of help. The ranks of IT professionals is expected to grow only 50% by the end of the decade. Yes, you'll be popular, but you won't have a lot of time to enjoy it.
And, if you're hoping for some magic technology, you might want to consider these two lines. One shows the steep declines associated with the technology (that's the red line). Amazing cost reductions, no?
Now consider the blue line -- which shows the aggregate spend increasing substantially as prices fall. Storage is getting ever-cheaper, and we're spending ever-more on it.
The intellectual rationale for organizing around a storage-as-a-service model is devastatingly simple. On the supply side, create standardized storage services and ruthlessly manage their efficiency. On the demand side, give people what they want and need -- without encouraging them to go elsewhere.
But that's not how things are generally done today, at least in many enterprise IT environments. And there's a very long list of perfectly valid reasons why this is the case.
But it will be the model -- it's only a matter of time as far as I can see.
What Makes A Good Storage Service?
Before we dig into the hard questions, let's step back and describe an idealized storage world.
The lofty goals of an enterprise storage-as-a-service model is simple:
(1) deliver the storage services people need and want
(2) using an easy-to-consume model
(3) competitively priced vs. external alternatives and
(4) with the baked-in protections the business needs
As always, the devil is in the details.
Total up your scores, and let's see how your storage operation stacks up, shall we?
Question 1 -- Do You Have A Separate Storage Team?
You'd be surprised how many decently-sized operations I meet that don't have a dedicated and well-organized storage team. Sure, they'll have a few people who know or do storage stuff, but it
won't be a formalized function. When these same people complain about their storage environment, it takes a sizable amount of willpower to avoid the inevitable eye-roll.
If you're going to deliver storage-as-a-service (or even manage storage effectively at scale), you're going to need a team that does that for a living.
- We have a few people who know storage (1-2 points)
- We have a few dedicated storage people (3-5 points)
- We have a storage team, but it's not exactly a well-defined function (6 points)
- We have a formal and established storage team with well-defined roles and responsibilities (7-10 points)
Question 2 -- Do You Have A Storage Service Catalog?
A "storage service catalog" is nothing more than a list of standard storage services that non-storage people can understand and consume. Besides the obvious benefits associated with standardization, it helps storage users understand their choices and costs associated.
No service catalog, no storage-as-a-service.
- We have a few options that we offer people -- different storage types, different protection types, etc. (1-3 points).
- We have a handful of standard storage service offerings that aren't described in technical terms (e.g. gold, silver, etc.) (4-6 points)
- We have a pretty well-developed set of standard storage service offerings, but people tend not to use them. (7-8 points)
- We have a well-developed set of standard storage service offerings, and they're used most of the time. (9-10 points).
Question 3 -- Is Your Storage Service Catalog Easy To Consume?
If I work elsewhere in IT, or perhaps somewhere in the business -- how easy is it to discover, request and receive a storage service?
Is it easy to find what services are available, and how to use them? How much effort -- as a user -- is required to actually get the service? Or are there poorly documented processes, lots of meetings, dozens of approvals, etc. etc.
Things that are hard to consume, well, don't get consumed much ...
- I'm not quite sure how someone would go about finding out about our storage capabilities, let alone request them. (1-3 points)
- We get lots of requests, but can't turn them around quickly or efficiently. (4-5 points).
- We can turn requests around pretty quickly, but people generally don't know about our capabilities and processes. (6-7 points)
- People generally know about our capabilities, and how to use them effectively (8-10 points)
Question 4 -- Do You Know Your Cost-To-Serve?
Every service provider knows their cost-to-serve: it's the all-in calculation of both fixed and variable costs associated with delivering a unit of customer service. However, when I speak with enterprise IT organizations, while they seem to know the costs of the ingredients (hardware and software), they are often far less precise on the cost to actually provision and deliver a specific end-user service.
Going farther, you can't really get into showback or chargeback until you have a good handle on your cost-to-serve.
- We sort of know what we're paying for storage, but not much else. (1-3 points)
- We have a rough idea of how to gross up our raw storage costs to approximate our cost-to-serve. (4-5 points)
- We have a decent model for cost-to-serve, but it doesn't capture a number of important factors. (6-8 points)
- We have a good model for cost-to-serve, and can explain it to the business when asked. (9-10 points)
Question 5 -- Do Your Internal Customers Know What Services They Are Consuming?
You can't deliver a service without some sort of bill or statement -- even if it's only a showback environment. If I'm a customer of your service, I'd really like to know a bunch of things: how much I'm consuming, associated costs, what the trends might be, any interesting options to reduce costs or improve functionality, are you meeting performance and availability goals, and so on.
Even better if all of that is in an easy-to-understand, easy-to-consume format.
- Our reporting capabilities are extremely limited: the information is hard to get to, and even harder to report back on an ongoing basis. (1-3 points)
- We're running some simple reports, but we generally don't show them to users unless asked. (4-5 points)
- We've invested in tools and processes that report back on a by-user and by-service basis. (6-8 points)
- We've invested in making that information easy to consume by the people who use the service. (9-10 points)
Question 6 -- Do You Measure Yourself As A Service Provider Would?
Service providers are in a business. Sure, they measure things like revenue and profitability -- but the better ones focus on things like market share, customer satisfaction and repeat business. If you're going to compete against external alternatives -- yes, cost matters -- but so do a bunch of other things that drive customer usage.
For example, are you easy to do business with? Are your rates competitive? Can you help me figure out what's best for me?
- Our internal measurement system mostly focuses on traditional IT concerns like resource costs, capacity utilization and uptime. (1-4 points)
- Our internal measurement system has time-to-provision a new or changed service as a key metric. (5-6 points)
- Our internal measurement system has a "customer satisfaction" component. (7-8 points)
- Our internal measurement system compares our services with the best-of-breed external offerings, and we share those comparisons with users. (9-10 points)
Question 7 -- Do You Have Well-Defined Continual Improvement Processes?
The services your users want aren't static entities, nor are the mechanisms you use to deliver them. Continual process improvement is the order of the day -- both at the front end of the service (what your users see) as well as the back end (how you deliver those services).
The goal? More services, better services -- and at continually declining costs.
But that won't happen unless you make it a formal process: something IT service providers understand well; enterprise IT groups less so.
- We don't seem to spend all that much time on either process or technology improvement. (1-3 points)
- We spend some time looking for better technology, but don't seem to spend much time on process improvement. (4-5 points)
- We spend considerable time on back-end process improvement. (6-7 points)
- We spend considerable time on front-end "user requirements and user experience" improvement. (8-10 points)
Question 8 -- Have You Changed Your Funding Model For Storage?
Service providers have a competitive advantage: when demand exceeds supply, they can go out and simply buy more capacity. Not the case for so many enterprise IT organizations, is it?
But unless IT organizations can quickly fulfill legitimate demand, their internal customers are likely to start looking elsewhere for better answers.
How enterprises approach IT finance is curious to me. So many finance organizations are comfortable with variable funding models with certain parts of the business (e.g. travel, headcount, raw materials, etc.) but are unwilling to apply this same line of thinking to IT consumption, which has a distinctly variable component to it -- or should!
IT is then put in the extremely disadvantaged position of having to both encourage and limit demand at the same time. I haven't met a service provider yet who tries to limit consumption.
Ideally, IT provides the services, and finance is in charge of constraining resource demand -- and not IT!
But there is hope: a natural progression arises from establishing service catalogs to enabling showback to creating the case for variable funding models. Unfortunately, until at least a portion of funding model has a variable component, you'll often end up hamstrung in your ability to meet your user's requirements.
- We get a fixed allocation, and we make do the best we can. (1-2 points)
- If demand exceeds supply, we make the requester go to finance to get us more funding. (3-4 points)
- If demand exceeds supply, we work with the finance people to see if we can get more funding. (5-6 points)
- If demand exceeds supply, that's sort of expected, and our finance people know how to deal with it. (7-8 points)
- IT is in charge of supplying services, finance is in charge of controlling and managing demand -- not us. (9-10).
Question 9 -- Is There Someone Doing Business Operations On Your Behalf?
If you think about it, a good-sized storage services operation is really a small business, and thus needs a bit of the same "business operations" skill set that any service provider would need.
Are our prices and costs in-line? Do we have enough capacity on the floor to meet immediate demands? Are we reaching our internal customers, and meeting their needs? Do our storage services offer the level of "trust" that the business demands?
Good business ops means you're running a good "business": happy customers, decent growth, management is happy, etc. Without business ops, you're just sort of keeping the technology humming.
- We don't do much of any of this, really. (1-2 points)
- We've got some help in costing and vendor negotiations, but that's about it. (3-4 points)
- We've got some help in capacity planning and forecasting that's used as input to the budgeting process. (5-6 points)
- We've got some help in figuring out how to price our services internally. (7-8 points)
- We've got a solid business ops capability: costs, prices, capacity, market share, trust requirements, etc. (9-10 points)
Question 10 -- Is Your Storage Team A Great Place To Work?
That's a hard question, but it turns out to be an important one. If your storage team is motivated to bring their best each and every day, you're doing something right. If not, you've got some serious challenges ahead.
Good storage talent is in demand -- and you're either attracting it, or not.
I get to meet happy, productive and progressive storage teams as part of my work here. I also get to meet plenty of the other kind as well. The difference is night and day.
- We seem to be continually challenged with attracting and retaining storage talent. (1-2 points)
- We've got one or two really good people, but if they decide to leave, we've got a major problem. (3-4 points)
- Morale in the storage group is good, and seems to be improving. (5-6 points)
- The storage team is amongst the internal "rock stars" of the IT group. (7-8 points)
- The storage team is often asked to speak externally on what they've done, and how they did it. (9-10 points)
It's Reckoning Time
Add up all your points, and let's see where you stand up in this informal survey ...
10 - 40 points:
A rather challenging situation, especially if you've got decent size and decent growth -- hopefully you've got undemanding users! For one thing, several external storage services would probably stack up better than what you're doing today. My harsh recommendation? You either need to make a substantial commitment to improving the situation, or perhaps consider a managed storage service for a while.
40 - 60 points:
Take heart! Some of the most important foundations you'll need in the future are already in place. Whether you were smart, or just lucky, you've got something good to work with. With a little focus and effort, you could completely reposition how your organization uses storage -- and compete successfully against external alternatives.
60 - 80 points:
You are well on your way to offering competitive storage services that your users want and the business will appreciate. There's still a good chunk of work to get to an idealized state (mostly outside of IT!), but all of the hard work is largely behind you now.
You've arrived! Your management team should have a portrait of the storage team in the lobby as a small token of their appreciation. Seriously, though, you're in the rare upper echelons of proficient enterprise storage teams -- or you work for a service provider! You should be giving serious thought to documenting your story and sharing with others. Yes, you're that good.
Back To The Beginning?
We're entering an information economy. Businesses are trying to construct their digital business models that will power them for the coming decades. Big data is becoming a powerful tool in its own right.
What does this mean? Lots of information, and consequently lots of storage. More than most of us can imagine. Sooner or later, so many IT organizations will be forced to sit down as a management team, and decide to get very serious on how they approach the whole topic.
Of all the potential models out there, the storage-as-a-service model has shown itself to be the most useful one for so many logical reasons. However, that's not how most enterprise IT organizations are doing it today -- but will need to before long.
And -- for all you storage professionals out there -- time to start thinking ahead ...