Thinking Outside the Bowl - Storage Expo 2005

2008-12-11

Hello and welcome to the second day of Storage Expo. Was anyone here today that was also here yesterday? Yes? As you know, industry analysts need to categorise things, and that put you in the category of people who need to get out more. Come on, this is storage!

But seriously (OK, I was being serious…), I’m not sure it’s particularly my place to stand up here and preach. The industry analyst's job is to understand what is going on in technology, and to help end user companies make informed decisions about how to use such technologies. I’ll be absolutely frank with you – I don’t believe that we’ve done that good a job of this. We used to – when it was down to things like databases or office automation packages or hard disk arrays, and it was quite straightforward to compare and contrast the features, or determine who was buying what.

Even as technology started to get more complicated, we analysts continued to provide a useful service – to cut through the complexity and explain things as they really were. A few years ago however we started getting a little too wowed by the marketing - everyone should have enterprise applications, we said, and all the enterprises bought the. Wow, that was clever, we thought, so we did it all over again with other things, like SANs. OK it wasn't just us saying it; then the Internet arrived and all hell broke loose. "E-business or no business," that was the mantra, sounded great but the trouble was it was just plain wrong. Sorry. There was a technological storm, where wave after wave of new concepts, gadgets, types of software, appliances and so on struck the shores of end user organisations, who were trying to take it all on at the same time as coping with the frustrations that were caused.

Rather than trying to help people cope, or attempting to slow things down in any way, the industry analyst positively encouraged this situation. Marketing models were developed for IT vendors to make the most of the situation – nothing wrong with that in principle, but let’s face it, the crossing the chasm model was designed specifically to help IT vendor companies sell more stuff. End users didn’t help – I’ll never forget a CIO show me a copy of “Management Today”, open on a one-page article written about the latest and greatest technology with a circle of highlighter pen and the words “When can we get one of these.” I think it was CRM, but I’m not too sure.

The crossing the chasm model – does everyone know what it is? Starts here with the early adopters, before dropping into the “chasm” and emerging as a mainstream technology. Over here … we have the laggards.

Trouble is – and I hate to break it to you guys – this curve is open to abuse. I know, I know, that’s terrible – but it’s true. Marketing departments the world over have not only been trying to second guess the curve but… would you credit this… they’ve been hyping up technologies to force them into the mainstream!

Impossible to believe, I know. Until suddenly, quite suddenly, a couple of years ago the end user community stopped listening. I think the last technology trying to get through the door as it slammed was location-based services, and weren’t they going to be great? The people I feel the most sorry for are the pizza companies – we were going to be able to find the nearest pizza place, anywhere in the world, their sales would have gone through the roof…

Since then we've had the bubble bursting, the downturn (never call it a recession) in our industry, the rapid deflation of technology share prices and frantic scrambling for some companies to keep in business. Those enterprise applications haven't turned mediocre businesses into global success stories, we’ve discovered what we knew in our hearts all along – technology alone can’t solve the problem. While technology can be part of the solution, there’s no such thing as a “technology solution”.

With all of this in mind, where are we now, now that the dust has settled? While there’s certainly a lot more positivity in the industry, the dynamic is different. We’re seeing a lot of consolidation work, many companies are re-organising their server and storage infrastructures and using them as a common basis for their applications. We’re seeing companies looking to extend their existing applications, for example to add collaboration facilities or better integrate them together. We’re seeing certain types of organisation in Europe building compliance features into their architectures – notably in the financial sector, or companies with strong linkages to the US. Indeed, the majority of today’s activity in larger companies involves improving what is already there, either by replacing it or extending it.

What does all this mean? We can see that the bulk of current IT activity is architectural in nature. Companies largely have the applications and services they need, and now they just want them to work better: surely, this is not too much to ask! Trouble is, current buying behaviour does not match any particular chasm model – as companies are trying to improve their lot rather than attempt some new way of doing things, it could be said that the chasm is already crossed for both infrastructure and for major applications, which let’s face it, together make up the majority of most organisation’s IT. We’re 80% there – now we’re just trying to work out the 20% to get things working together as well as possible.

Indeed, there are plenty of technologies that have well and truly crossed the chasm. ASP’s, for example – that’s application service providers – were another casualty of the bubble bursting. They’ve been quietly getting on with the job however, and are now well and truly mainstream. Consider this picture for example – can anyone tell me what it is? Now, if anyone can tell me where it is, then that deserves a prize! This is World of Warcraft, a massively multiplayer online role playing game, or MMORPG. You may think that showing this is no more than a thinly veiled attempt by me to turn a computer game into a taxable expense, and you’d be… no, I wouldn’t be so shallow. (nod) (shake) (nod) etc

World of Warcraft is also an ASP. Not only is it hugely popular, but it’s also virtually impossible to pirate, as it follows a subscription based model. And it works – extremely well. Many of the reasons cited for ASP failure, which were largely around service quality, have had to be solved for this to become a reality. Not only is it a growing phenomenon, it is proof if we need it that the chasm has already been crossed for infrastructure. The role playing genre may well be inside a tornado of its own, but that’s another story! Did you know there are now currency markets for virtual gold?

But we digress. While we’re on the subject of virtual things, this is possibly an appropriate moment to mention storage virtualisation. What do we mean by storage virtualisation? Essentially, virtualisation offers a layer, through software or through a hardware appliance, so that we can manage and provision our storage resources as though they were one, big, “virtual” pool. Virtualisation can exist in a number of places in the infrastructure, and I’m not going to go into the technical detail here. There is a hall full of people out there that have been girding their loins to talk to you about exactly that for a number of weeks now. What I will tell you – and this will probably spoil their pitch – is that virtualisation is not a product, not as such. There may be vendor companies that sell it, no end user company will ever be using it in isolation.

Of course, I don’t really need to tell you this, as you already know. However, someone does need to tell some of the vendors. Virtualisation cannot cross the chasm because it is an architectural construct. We already employ virtualisation techniques in a variety of ways on computers and around the network, we have virtual memory, virtual LANs and so on. The ability to virtualise storage is no more or less than the storage vendors catching up, and implementing open mechanisms so that their storage hardware can be accessed transparently. Rather than greeting virtualisation with delight, many end user companies are reacting with relief – “at last, you’re giving is what we needed all along.”

These architectural mechanisms cannot cross chasms, instead, they are steps along the way for a better managed, more efficient infrastructure. I call this the “Arathi” model, if anyone’s interested – you heard it here first – as you can see we are on the way up the mountain, but there is no single peak. If we were honest with ourselves, we would see technologies such as virtualisation as way points on our journey towards better storage , but instead we insist on hyping them up, treating them as goals in themselves. As a result, they appear as a series of false summits. Virtualisation is one of these – it may have certain benefits, Nothing is more exhausting or debilitating to a climber than a false summit.

OK, that’s virtualisation. What of ILM in all of this? Infrastructure Lifecycle Management, ILM, another term that is bandied about. What’s that all about then? Let’s get one thing straight. You can’t buy ILM. Its absolutely not a product, you’ll never see it in a catalogue or printed on a CD-ROM. You’ll never install it, or patch it, or debug it. So what exactly is it?

Let me put it this way. If virtualisation is a false summit, then for storage, ILM is the mountain. It’s not about hierarchical storage management and tiered storage, or efficient archiving, or full-index search mechanisms, each of which are just facets of a highly efficient, well-managed storage architecture. You can’t have one without the other – the high efficiency and the good management.

While from an end user perspective we might be looking at mountains, chasms and all that, from an end user company perspective I believe the model should look more like waves. This is a series of waves hitting the Great Barrier Reef, so you know.

There are, essentially three phases that we need to go through to reach the top, none of which can happen in isolation, but some of which are dependent on the others. There are essentially three, distinct phases that lead us towards the goals of ILM, and they need to be followed in order. Or rather, they will be followed in order, whatever we would like to think or our IT suppliers would like to think, for the simple reason that the dependencies between them go in one direction only.

What are these phases? Here we go. We have infrastructure consolidation, followed by resource management, and then service management. Infrastructure consolidation comes first, because you need something to manage after all. Of course it could be argued, one could get on and manage the existing, convoluted legacy environments, but you seem to have made the decision despite what us “experts” might say – you want to consolidate first, thank you very much.

Second comes resource management, or making the most of what you’ve got. Virtualisation fits here – it’s software designed to help hardware use more efficient – but hang on, haven’t we a name for that already? We have – its called “the operating system”. That’s all we’re doing here – reinventing the wheel. Sure, it’s a much bigger, globally distributed wheel, but it’s a wheel nonetheless. At the moment, all this clever technology for virtualisation, for hardware orchestration, for intelligent archiving and so on, all of it exists as isolated packages but they’re currently being integrated together. In five years’ time, its not that you won’t even recognise the packages, you won’t even see them!

Thirdly we have service management, which depends on the ability to manage resources. This is about delivering a service to the business – in other words, when real end-users need something, they get it, and they are held accountable for it. This is about resource allocation, SLA management, all these terms, essentially it is outward facing towards the business applications and their users.

Another way of looking at this is shown here. Did you see what I did there? From the business perspective, first, we need to get the basic hardware platform in place, then we need to deliver on how we operate these platforms, for our own benefit as IT people running an efficient operation – we may be able to automate certain aspects of these processes, using tools such as virtualisation software. Only then are we really going to be able to deliver an optimal service to our end users – the business – the people. In storage terms, achievement of all three equates to achievement of something resembling ILM, by any definition of the term. Should it include content management? Of course! Indeed ILM without content management isn’t really worth having, as it lacks the linkages into business information flows, which let’s face it, are pretty fundamental. You can’t know what to do with your data unless you know what its for.

OK so far? Right, what I wanted to do was pick up on some proof points of what I’m saying, from our research. There are a stack of studies I could have drawn on, I would stress if you have any questions at all or if you want to debate anything, please do get in touch, my email is at the end of this presentation. I should also mention, you can sign up for our reports for free, no obligation, just mail me or check the Quocirca web site if you’re interested.

Let’s put some meat on the bones, then. First, infrastructure consolidation: as we mentioned before, this is well and truly underway, as shown here by some research we conducted over the summer for EMC. Here it is – you can pick up a copy from the EMC stand. We conducted another survey for Oracle recently, on grid and virtualisation, and we had similar results. That’s about 90% that plan to consolidate, and two thirds that have projects underway – fair to say there’s a trend there.

When we look at software technologies to make the most of such consolidated infrastructures, we don’t get the impression that things are up and running just yet. A third of companies see virtualisation as majorly important, and most of the rest give it some credit, but its hardly a glowing endorsement, is it? Other research we’ve done backs this up – 60% of companies are seeing virtualisation as an option, and 40% are doing something with it, but these numbers are growing. It looks like a second wave trend to me.

As we said before though, all this clever stuff is only really worth doing if we have the right processes in place to take advantage of it. We haven’t – given the fact that a third of companies aren’t even doing backups correctly, if at all, what chance do we stand with higher level processes such as managing a virtualised storage environment? To go back to the mountaineering analogy, its like putting on crampons without wearing any shoes. The only results will be sore feet and blisters.

This is an important point. I was a bit harsh on vendors earlier perhaps, suggesting that they are just opportunist sales folks, taking advantage of an unsuspecting public. Surely not… but the truth is, they wouldn’t do so if we didn’t let ‘em. Let’s face it, we’d all love to believe that the marketing was true, even today, that our problems really would all be solved. We also all know that if we put our own houses in order, we’d be better able to serve our own companies. There are always reasons a-plenty – conflicting priorities, shortage of time and so on, all of which are valid, but the point remains – nobody should expect to be able to run a super-efficient storage infrastructure without putting the operational basics in place. It stands to reason.

To move on to the third circle, that of service management, I have a silly question for you: should we want to do this stuff anyway? As a question, it’s a trifle unfair – which CIO is going to say he or she wants to run an inefficient shop, totally ignoring business need as he does just what he wants? Sure enough, when we ask this kind of question, we get a unanimous, positive response. So, we don’t ask it very often. What we do ask is how important are certain types of technology to the business – email is the rising star, as a study we performed for Dell in July, demonstrated.

Email was very important, sure, as you can see – 90% of respondents thought so. When we asked how many sales actually were conducted via email, the response came back as 25% - that’s a heck of a lot. Indeed, it’s a bit scary – I’m not sure Exchange or Notes were designed as the ultra-resilient platform for 25% of the world’s sales transactions! And that’s without considering that the underlying servers and storage are up to the job.

Its just one indication of how important it is to get IT right for the business. Another example, based on our more recent research, was when we asked questions about what people would like to see improved in their existing IT. All of the questions were answered yes, as you might expect, but most interestingly, “Finding information” is top of the list of all issues. If that’s not a business issue, I don’t know what is – and we know this, we know through the time we waste ourselves, hunting for that email or file, matching up one customer record with another, rifling through our pockets for a business card and so on. This is not a problem to be solved with some clever technology, rather we need to be better organised about what we are storing, where and why. In short, this is a people problem.

So, we cannot succeed through technology alone. This is the bit where I start wrapping up – three more slides and we’re done. The way we see this is as the law of diminishing returns – it gets harder as we move through the phases. It is no wonder that we concentrate on buying new technology, because that’s something we know how to do – I used to run a network, I was that soldier, I know. Far harder is to understand how to improve how we do things, build better relationships with our internal corporate customers, particularly if there is no impetus from board level. This guy – Thomas Malthus – understood first that there are only so many resources to go around, and the more you dig into the resources, the harder it becomes to support the problem space as it grows, be it people, complexity, relationships, data quantities or whatever.

Given all of this, we’ve had a fair crack at trying to establish where exactly we are along the road to achieving these goals. Or should I say, how high up the beach? We asked a whole set of questions to cover each of these phases, and used the responses to generate some kind of index. Fascinatingly, as you can see, while the individual questions and answers were totally different for each of these sub-indices, the results have come out awfully similar – overall then, in the UK and Ireland, we can say that similar levels of progress have been made across all three phases. These are marks out of ten by the way – looks like we’re just over half way, but let’s not forget the law of diminishing returns!

Of course the index values by themselves mean very little; however it does give us a starting point for comparisons. Here’s one example, by industry – the only point I want to draw out, is that “other public sector”, by which we mean, non-healthcare, that’s government departments, councils and so on, is definitely falling behind, whereas utility companies are edging ahead. We can drill down and see why this is, I’m not suggesting we do this now but if you want further information you can pick up a summary of the report from the EMC stand.

So, what was all that about “thinking outside the bowl”? Weve had chasms, and bridges, mountains and false summits, journeys and waypoints. The point is, we’re not going to get there by thinking about technology alone. The IT departments of many companies realise this already, and the ones that use their business requirements to drive how they implement their IT, and who implement comprehensive operational processes to boot, will stand more chance of “getting there” than those that rely on technology alone. We do not need vendors, or anybody else for the at matter, to tell us what technologies we should be considering, be they dressed up in terms such as virtualisation, ILM or whatever. What we do need are companies that can work with us to understand the needs of our own businesses, and help us to define and deploy technologies that work for us, not for them.

With that in mind, it remains for me to thank you for listening, and to say I hope you enjoy the rest of the conference.