When I announced that I was leaving EMC and VMware to join Oracle, the reaction was interesting to say the least.
There was a certain crowd that thought I’d be part of the EMC Federation forever. There was another crowd that couldn’t truly believe I’d join Oracle, and enthusiastically at that.
Gratifyingly, there were also plenty of well-wishers -- that was nice!
At the time, I really couldn’t go into the reasons why this made such logical sense to me. It had a lot to do with what I was seeing in enterprise IT – both on the supply and demand side.
When the world changes, you need to change as well.
So I did.
Obligatory disclaimer: these are entirely my personal views – they have not been reviewed nor approved by my employer. I take full personal responsibility for everything I say here.
A Jaundiced View Of Enterprise IT Organizations
I’ve enjoyed studying literally thousands of IT organizations from multiple angles. The more I look, the more new patterns emerge. I find it fascinating. I guess I’m easily entertained.
One interesting lens is segmentation by the value of the applications they are responsible for. By value, I mean “value to the people paying the bills”.
Order processing, for example, generates revenue for a company, so it’s pretty close to the top of the enterprise value pyramid. VDI, while nice, is more of an optimization than a money-maker for the company, so it’s typically somewhat lower down in the food chain.
My belief is that – when resources are limited – IT organizations will tend to invest in the applications that are of high value to the enterprise, and tend to starve the others.
Through that admittedly jaundiced view “what’s valuable to the business?”, so much of what we talk about in enterprise IT quickly becomes less interesting.
For example, hypervisors are certainly cool, but they essentially save money; they don’t make money. Containers are also cool, but only if they help bring new revenue-generating applications to market faster. Otherwise, they’re just a distraction.
A flash array lashed to a critical database can do more work, potentially making money – but flash arrays are quickly losing meaningful differentiation, and are now competing on price. At the end of the day, flash is flash -- unless someone can come up with some real secret sauce.
Show me an application or technology that clearly makes money (or avoids serious risks) for a company that uses it, and I’ll show you something enterprise IT is willing to invest in – ahead of everything else.
A Similarly Jaundiced View Of Enterprise IT Vendors
There is a black hole in our enterprise IT vendor universe: and it’s called commoditization.
Whether we’re talking on-premises technology, public clouds or both – everyone is slowly yet inevitably getting pulled in.
Even the over-marketed startups.
How many tech pitches do we see about “saving money?” Not that saving money isn’t important, but you can’t save your way to prosperity.
How many tech firms are routinely laying off staff, writing off dodgy acquisitions, starting to pay dividends, aggressively buying back shares, dealing with activist investors, struggling with IPOs, etc.?
None of these are signs of growth.
I believe customers win with commoditization in the short term, but not so much in the long term.
Yes, prices go down, but so do vendor margins. And a good portion of those margins get reinvested in R&D – making better stuff. Commoditization is like kryptonite to innovation. It gets harder for a large company to innovate in any form when it’s getting sucked into the black hole of commoditization.
Back To Enterprise IT Shops
That same black hole of commoditization is also affecting enterprise IT organizations, but in different ways. Many are fashioning themselves to look like internal IT service providers, and benchmarking their services against external alternatives. They have no choice in the matter.
Budgets are tight, and CIOs routinely tell me that finding (and paying for) good people is getting increasingly harder – even in the really big shops. Face it: a career in enterprise IT isn’t as glamorous as it once was, and many bright young graduates are choosing different paths.
These enterprise IT shops would rely more on vendor expertise, but it’s now a pay-for-what-you-get model, thanks to commoditization. So more contractors and consultants are used, hollowing out the core of enterprise IT knowledge even further.
"Science projects" are most decidedly becoming less fashionable. There's no money, no time and scarce talents. IT shops now want complete solutions that deliver proven an quantifiable value, so they can focus on more important things.
It’s not pretty on either side of the fence.
What I Concluded
If you’re an enterprise IT vendor and you only sell horizontal technology (e.g. stuff that can be used anywhere for anything), you’re heading into the black hole relatively quickly.
Witness the server vendors, storage vendors and network vendors. They are all attempting to reinvent themselves -- as what, I don’t know. I do think VMware is in a far better position than most, but will still have to paddle awfully fast in the coming years.
No, applications make money for the enterprise, and unless – as a vendor -- you’ve got a solid base of revenue-generating applications and supporting technologies that directly and visibly help, gravity will suck you in.
I believe that the more complete and integrated the solution a vendor can offer, the more resistant they will be.
And then there’s The Cloud in all its sundry forms.
Unless you can show your customers a wide range of complementary cloud services to go with your applications and supported technologies, you as a vendor will be sucked in even faster.
No compatible cloud option? Back of the line for you …
HP doesn’t appear to have much of a cloud that offers compatible consumption options that mirror what's sold on-premises. Nor does Dell. Or Cisco, or EMC/VCE, or NetApp, and IBM’s cloud doesn’t look particularly healthy to me. I think Microsoft has a reasonable cloud strategy, because it complements their “applications and technologies” portfolio quite well – but it doesn't appear to have chance of addressing critical applications in decent-sized enterprises.
Now, let’s juxtapose what’s going on in enterprise IT.
Fewer shops can justify building IT infrastructure on their own. It’s expensive, it’s time-consuming, and it’s risky. Especially when it comes to the infrastructure supporting those revenue-generating applications that keep the company running.
Witness the rise of reference architectures, converged systems, hyperconverged systems and their ilk.
But none of these are really integrated with the important databases and applications. To be specific, they are not tightly integrated with those parts of IT that actually make money. So these newer animals end up being a cost-savings play for their customers; and not so much a revenue-generation play.
Nice, but gravity sucks, doesn’t it?
And how many of these reference architectures, converged systems, hyperconverged systems, etc. – have a precisely compatible consumption option in the public cloud?
None – as far as I know.
As a result, I believe that they too will be drawn into the black hole of commoditization – it’s only a matter of time.
People will pay for things that help make them money before they pay for things that save them money.
When Demand And Supply Meet
Enterprise IT organizations are getting hollowed out. Most will have to prioritize.
At the top of the list, those critical applications that are essential to the business: taking orders, delivering product, supporting customers, analytics, etc. Make them run faster, make them run better, secure them, remove risks, etc.
The other stuff will have to take a back seat.
These shops aren’t going to have the luxury of hand-crafting critical IT infrastructure for much longer. It’s too hard, and too important.
They will look at cloud options at the periphery for now, but over time more critical enterprise apps will want to go there.
I believe they will prefer to work with suppliers who understand and can deliver applications, databases and the infrastructure that makes it all hum along – offered as complete engineered solutions, not as science projects.
They will prefer having precisely compatible cloud consumption options – perhaps more as an attractive option than an immediate need.
The few vendors that can pull this off will have a powerful form of anti-gravity – they will be highly resistant to the forces of commoditization that are sucking so many others downward.
They will charge a justifiable premium for the value they create. That premium can be reinvested in creating better answers to enterprise IT challenges.
Gravity can be overcome.
And, in a nutshell, that’s why I joined Oracle.
-------------------
Like this post? Why not subscribe via email?
The product of Gilder's Telecosm, and Carr's The Big Switch will yield similar conclusions. Compute, Storage and Network is a commodity. The value remains: Apps, Data, and Identities (of entities) with Access, so long as they can be de-coupled from (any particular) Infrastructure.
Posted by: Rawipfel | September 14, 2015 at 11:00 PM
Hi Chuck,
First of all, I wish you well at your new job. Hope they do not 'corrupt' your free thinking and candid opinions. Keep them coming.
I really like you analogy about commoditization and a black hole. And I fully agree. It slowly sucks everything to its center and you have to keep paddling hard to stay away. And technology is so fluid that the pull into the black hole is getting stronger every year.
Hardware vendors are already struggling. As you said, flash is just flash. Now some flash is better (faster, more reliable) than other flash but that is just a matter of time. And unless you can develop a game changing piece of hardware, you are going to get sucked into the hole.
There is more room in differentiation in the IT software and services world. But still, too many service providers want to be another Amazon. That is not going to work. The need to become something different, compelling. Same with software, you need to stay ahead of the pack.
At the infrastructure layer that is more difficult than at the application layer where you can directly influence business. And Oracle plays a big role in that space.
I may be wrong, but I associate Oracle with the more traditional IT (2nd platform if you will). And don't get me wrong, probably 80-90% of IT is still there. So a huge market.
Not sure how many startup disrupters use Oracles technology. Specially looking at integrated solutions (Exadata, etc). They probably do use Java (or the Spring variant).
And I still have fundamental issues with creating a whole forest of silos of different integrated solutions.
Having an Exadata from Oracle, next to a PureApp from IBM, etc. etc. and on top of that a generic platform for the rest. I still believe the best solution is to create a common Infrastructure platform that will run all your workloads. The advantage of tighter integration within a single stack does not out-way the simplification (and therefor automation) of the common infrastructure layer.
And that is not just driven by cost, but by the fundamental law that complexity of multiple silos increases risk and thus reduces quality, control, security, continuity, compliance, etc.
There are interesting times ahead :)
Stay well and keep your opinions coming...
Posted by: Aernoud van de Graaff | September 15, 2015 at 04:22 AM
Chuck,
First, congratulations on the move. I was one of the ones who probably thought you would be with the federation forever, but I agree with your motivation 100%. I am a long time fan of your blog and glad you’re still writing. I remember being in a room with Joe Tucci in 2010 and I ‘accidently’ used the “c-word” (commoditization), man the look I got – then Joe proceeded to say that there was absolutely no commoditization at EMC.
I left that year and joined a company called Storwize, which sold a “product” for doing real-time compression. That company was bought by IBM and now the “product” is a “feature” in their Storwize (they took the name for their new storage platform) / SVC storage solutions. I think we have seen this time and time again. Vendors that are being squeezed can invest in R&D as you called out so they acquire small startup companies with great ideas, spend some money and integrate the product into a solution to try to add value for its clients.
I also agree with you that “…flash is flash -- unless someone can come up with some real secret sauce.” The question here is what is that “secret sauce”. From your post it would be something that enables flash to help enable money making solutions for the customer. I’d like your perspective. This whole new space called “Copy Data Management - CDM” is taking on a life of its own. While today it is predominately used to solve data protection problems (seems like every IT shop has budget for that – back to your ‘risk’ comment), the reality is, the right CDM solution that can take advantage of the storage services on the flash array that the client has already bought and paid for can be much more powerful than just protection. If a company’s competitive advantage comes from the ability to deliver goods and services faster than the other guy, then a DevOps model for companies is a must in order to out maneuver its competition. Flash is part of that, but automation and orchestration of the data is as much or even more important. In a post you did the other day, you said “-- information technology -- is really about information and much less about technology.” It is the information that will make a business money. Also in this piece you talk about “…CIOs routinely tell me that finding (and paying for) good people is getting increasingly harder…” Technologies like Chef and Puppet are helping to alleviate some of this but in the storage world there aren’t a ton of solutions. The ability to drive automation (IT best practices) on a platform that can help development and test deliver solutions faster can help a company make more money, while at the same time, help them save money, not necessarily on the data copies (while it can) but on the labor surrounding the management of that data.
I wonder, would that be some of the ‘secrete sauce’ you were talking about?
One other thought. On your cloud thought, I agree, it is high time that people stop thinking of the cloud as just ‘cheap and deep’ storage. Leveraging the cloud for its economics when it comes to the ability to spin applications up in the cloud, perform the functions you need and spin them down, again saves money. But, as you put it, gravity still applies. Getting data in and out of the cloud is still something that folks need to contend with no matter how good the applications are that reside there. I think a data management solution that can help IT perform the same automation and orchestration around data management is powerful in helping to better exploit the applications that run well in the cloud. (BTW, my opinion only, IBM’s cloud with SoftLayer does seem to be pretty powerful, especially if having a “complementary set of cloud services and applications” is important, I think they have a good solution, but that is my $.02)
Best of luck to you Chuck at Oracle. I look forward to more of your enlighten reading.
Best,
The Storage Alchemist
Posted by: Steve | September 15, 2015 at 09:57 AM