Yes, Something Very New Has Come Along and more...

Yes, Something Very New Has Come Along

BrightstarIf you're a grizzled infrastructure guy like me, you're completely justified in your skepticism when any vendor claims to have announced something Truly NewTM.  

I mean, how many flash arrays, converged thingies, etc. does the world need?

Because we tend to focus on the underlying technology, we tend to miss other equally important innovations.  For example, Uber didn't really introduce new technology to the world; they just changed a familiar consumption model.

A few weeks back, Oracle announced a new industry category -- cloud machines -- under the banner "Cloud At Customer". First up: the Oracle Cloud Machine -- on-prem PaaS/IaaS targeted at enterprise application developers.

Simply put, it's a public cloud model delivered in the data center.  It fundamentally changes the familiar consumption model.

In the short time since, I've been seriously stunned by the level of customer and partner interest. People  immediately grasp the concept, realize that it's fundamentally different alternative, and are immediately curious.

We must be on to something here :)

Sure, I personally thought the notion of a cloud machine was going to be successful.  But at the end of the day, what I think doesn't matter that much; the opinion of thousands of IT professionals matters far more.

And -- so far, so good.

Oracle Cloud Machine In 30 Seconds

30secThe Oracle Cloud Machine is a re-packaging of a portion of the public Oracle Cloud for on-premises use.

The user experience is the same, e.g. public Oracle Cloud.  The operational model is the same as the public Oracle Cloud, e.g. the service is operated and managed by the same team (and using largely the same processes) as the public Oracle Cloud. The financial model is the same as well, e.g. metered or subscription. And, of course, the software stack and user experience is identical.

Hence the description of "public cloud model delivered on-premises".

In addition to the current model targeted at enterprise application development and execution, expect two more versions before long  -- aimed at database and big data analytics. Once understood, you'll see that the model is easily extensible to other portions of the IT landscape.

Hence it's best to think about cloud machines as a new category -- they are that different.

How A Cloud Machine Is Different

DifferentFolks used to traditional on-premises private clouds may scratch their heads, and wonder why all the fuss? After all, many of the ingredients look the same to them: compute, hypervisor, PaaS, etc.

First, this is integrated PaaS/IaaS targeted at enterprise applications developers. Standalone IaaS doesn't do this unless you spend a lot of time and effort rolling your own, and few application teams want to wait for the science project to finish.

Second, you're contracting for a complete, integrated service provided by Oracle. You're not buying the separate ingredients and cooking your own meal. It's a pure opex services model, just like the public cloud.

Third, it is derived from -- and seamlessly interoperable with -- with the public Oracle Cloud. No homegrown solution can make that statement.

ControlFolks used to familiar public cloud services also do their own head-scratching, but the questions are different.

A cloud machine lives on the premises of your choosing, under IT control. By itself, it is not dramatically elastic, but -- working with the public Oracle Cloud -- it is.

Like most public cloud services, developers are free to bring many of their favorite open source tool sets. Like most public cloud services, the infrastructure team doesn't play a direct role in providing supporting services.

Unlike public cloud services, you don't have the usual bugaboos about not being in control, data leaving the premises, latency, etc..

So the developers are free to move as fast and confidently, which matters.

What Senior IT Leaders Tell Me They Like About Cloud Machines

The senior crowd clearly appreciates the pure opex model of a cloud machine. This isn't clever financing in disguise, it's a clearly-defined integrated service you commit to for a fixed period of time. All costs are included: transparently and known in advance.

And when you're done, the service disappears and that's the end of your commitment. Just like the public cloud.

Smiling_execThese same people also appreciate that the Oracle Cloud Machine service is pre-integrated, and supported as a whole. This is not some reference architecture or solution guide; it's a baked, standardized product offering. Less time, less risk, less cost. Not to mention that Oracle has clearly integrated the whole stack: from application to database to infrastructure to support and so on.

On a more strategic note, I see that they appreciate how it brings immediate cloud agility to their enterprise application portfolio. No need to wait patiently for the various teams to cobble something together. There's also the attractiveness of easily mixing and matching with identical services from the public Oracle Cloud.

I do find it, errr,  interesting that senior IT execs who are responsible for multiple disciplines immediately grasp the concept.  Not so with the usual industry analysts, who tend to hold a partitioned (and historical) view of the world. 

What Partners Like

Oracle partners tell me they like being able to offer their clients a pure passthrough opex model for an integrated solution that sits behind one or more enterprise applications: either IaaS only, or a mix of PaaS and IaaS. It certainly helps them stand out from the crowd -- something that every partner wants to do.

I mean, just about anyone could sell you the ingredients -- how many sellers can offer a public cloud model on-premises?

It also simplifies their partner business model -- they can focus on the application, and leave the supporting infrastructure and tools to Oracle. Everything they position can run on-premises, in the public Oracle Cloud, or any combination. That's very strong differentiation.

The Infrastructure Team -- Well, Not So Much

You can't please everyone.

UnhappyHey, guys, I've worked the infrastructure end of the business my whole career.

I'd love to say cloud machines are all about the infrastructure team and what you do, but the truth is completely the opposite: it's about delivering public cloud infrastructure services on premises without the need for direct involvement by the infrastructure team.

Yes, I know you're heavily invested in VMware and everything that goes with it. Not relevant in this context. Business users want results, and aren't as obsessed with the ingredients as you've been.

More importantly, they can't afford to wait for y'all to put something together on their behalf.  Even if you could.

I gently remind people that part of being an IT professional is continually evolving your skill set forward, especially during big industry transitions. That NetWare cert I got many years ago doesn't help me much these days, nor does my extensive COBOL programming experience :)

The industry will still demand skilled IT professionals in the future; cloud or no cloud. They'll just be working with a different model.  And the market for people who specialize in hand-crafting infrastructure appears to have started its inevitable decline.

Competitive Responses? There Aren't Any.

GotnothingI can safely assert that no IT vendor is delivering anything that looks like a cloud machine today. Going farther, it's hard for me to see how any one of them could do so in the near future.

They'd need a full application stack, an enterprise-oriented public cloud, as well as the proven ability to deliver enterprise-class solutions on-premises.

Rare is it indeed in our industry when one vendor can do something so attractive for so many IT shops, and no competitors can mount a meaningful response.

I did see one deck from a competitor decrying Oracle "lock in".

As I remind people, deep integration is the flip side of lock in. I drive a Ford F-150, which I love. I use Ford parts when things break. I don't feel locked in. Besides, users are free to run almost any application (Oracle or otherwise), bring their favorite tool sets, or move the software stack elsewhere if needed.  Let's move on, shall we?

Sometimes I think some IT folks who obsess about lock in might be looking for job security in the multi-vendor integration and support business. As the industry moves to more integrated stacks, that's going to be a losing battle.

The Bottom Line - Faster, Now!

FasterThere are so many organizations I meet these days that are near-desperate to get more agility from their enterprise application portfolio. The app team is working their butts off, but don't have the modern tools they need to work effectively.  

They look longingly at public clouds, and wish they could get the same on their premises.

To be fair, the infrastructure team is working their butts off as well, but just can't close the gap.  Even if they could replicate the technology model (very difficult), there's still the financial and operational model to clone.  Oh yes, and whatever you eventually end up building won't have a compatible public cloud option.

The Oracle Cloud Machine solves this thorny IT problem brilliantly.

And this is getting fun for me :)


Like this post?  Why not subscribe via email?


On Being Happy

Why_hereFrom time to time, I am preoccupied with a simple question: why are we all here?

Perhaps the same thought nags at you as well.

I prefer answers that are simple and direct, and don't require invoking a cosmological constant. It wasn't until later in life that I came up with an answer that works for me.

Your results may be different, naturally.

I believe it's very simple: we are here to be happy -- individually and collectively.

Learning To Be Happy

Happy_meI think the first mission we have as human beings is to learn how to make ourselves happy. As young children, it seems to come quite easily, but it gets more complicated as we become adults.

I think it has something to do with hormones.

As an adult, I was eventually successful in teaching myself what made me happy. It wasn't obvious at first. I lost a lot of time chasing things that other people told me were supposed to make me happy, but didn't. My world changed significantly for the better when I finally dialed in the working formula.

A key learning: listen to yourself, not others.

I often meet adults of all ages and all walks of life that haven't succeeded at this fundamental mission yet. I feel for them, and help as I can. May you have an epiphany as I did.

When you're happy, you tend to live a harmonious life. You don't quarrel, or have many bad days. You aren't easily irritated, and you don't rant much. Your energy and interactions tend to be positive, which engenders the same from others. People want to interact with you as a result.

And, most importantly, being happy at your core enables you to do more for others.

Helping Those Close To You To Be Happy

SculptureLike many of you, I am in a long-term married relationship with three wonderful kids. I really, really want them to be happy, much as I am. I work hard at helping them as needed, but it's fundamentally their job -- not mine.

My wife is very happy these days. She is doing exactly what she wants to do, and is quite good at it. Her days are long, but she enjoys it deeply. Seeing her so satisfied of course makes me happy.

Yes, it sounds dopey, but it's true -- you can't be happy unless your spouse/partner is happy as well. I have no problem putting her happiness ahead of mine as need be.

My three children are all happy young adults. They feel good about themselves, their place in life, and what the future will bring them. As a result, they tend to make good decisions, which is important.

Seeing them on the road to their future makes me deeply pleased.

Helping Others Be Happy

Once you've taken care of yourself and those close to you, you're in a stronger position to help others. I've found that's very hard to do if you're worrying about your own situation, or that of your family.

For example, I routinely make the time to encourage and mentor those that I work with. Skip the corporate mentoring program, and just do it because it feels good to help others. Most of the time, I help people discover what they really want out of work and their career. Often, they've adopted someone else's goals without going through their own internal discovery process. That's sort of like wearing someone else's clothes and expecting to be comfortable.

I give of my time and money. If someone I've met is going through a rough patch, I listen with empathy, offer encouragement and refrain from giving specific advice unless asked. When I retire, I look forward to investing more time in younger people going through their formative years -- it can be a rough patch for many.

Not everyone means well. If someone appears intent on putting a dent in my positive attitude, I steer clear and refrain from interaction. Lots of people out there working on lots of issues, and not all of them can be helped.

A Simple Philosophy

Thinking about the world this way makes it very easy in my interactions with others: I try to help them get what they want unless it would adversely impact someone. Happiness shouldn't be a zero sum game. When I can help someone out, I feel good -- it's a win-win.

So, what's your philosophy?



IBM And VMware Tie Up: What Does It Mean?

MatchOur IT industry has always adored "strategic relationships", we seem to read about several every week.  They can be convenient answers to painful gaps in individual strategies.  

Some work out, most don't.  Hope springs eternal.

But the announcement of IBM and VMware getting together to deliver cloud services certainly deserves closer inspection -- at least by me.

On one level, it makes perfect sense.  On another level, not so much.

While we'll likely have to wait a year or more to see if this particular spawning bears any offspring, it's definitely worth discussing.

What Was Announced

From the press release:

IBM and VMware have jointly designed an architecture and cloud offering that will enable customers to automatically provision pre-configured VMware SDDC environments, consisting of VMware vSphere, NSX and Virtual SAN on the IBM Cloud

 Another take from someone writing on Forbes is here.  And I'm sure there will be many more opinion pieces to join mine over the next few days :)

No real specifics were discussed about actual availability, as you'd expect.

Why This Makes Sense

On one level, you can see what each party gets from this arrangement.

Cloud_v_hybridVMware, which has been telling a hybrid cloud story for at least five years, has yet to deliver an attractive, compatible enterprise-class public cloud offering to complement their on-premises capabilities.  First, there was supposed to be vCloud Director and a partner network.  Without being critical, vCloud Air has struggled.  

And then there was that recent headfake around EMC and Virtustream.

Bottom line: VMware needs a compatible, enterprise-class cloud, and fast.  Smaller partners couldn't deliver.  VMware itself couldn't deliver.  EMC couldn't deliver.  Presumably, this wasn't a business that Dell was interested in.  HP is still smarting from cancelling their last big cloud initiative.  Cisco?  We won't go there.

So IBM is a logical choice.  Ain't many other choices out there, are there?

IBM theoretically gets access to an enormous vSphere installed base to sell compatible cloud services.  My joke about BlueMix is that it's like dark matter; I'm sure it's out there, I just don't have any compelling evidence.  IBM can potentially add standardized, easy-to-consume vSphere-based cloud services to their existing portfolio, and doesn't really give up anything in the process.

On paper, there's a certain logic to this.  Not mention urgency, as Amazon / Microsoft / Oracle all are in the cloud business today, so time is short.

What Doesn't Make Sense

The first thing that struck me is that VMware theoretically ends up doing quite well, IBM not as much.  

CrossedThe offering will be comprised of commodity hardware (low margin), VMware software products (margin goes to VMware), and IBM's value-added service delivery.  

I don't see much room for upsell for additional services, as these will be competing with customers who are quite comfortable today doing infrastructure in-house -- whether assembled or pre-integrated.

IBM can charge a small premium, but not much. 

The other invisible piece of the equation is the cloud  service provider operating system: that entity that provisions clusters and VMs on behalf of multiple tenants.  

This is a non-trivial piece of software to create and mature.  Will IBM take existing vCloud Air code and try and whip it into shape?  Create their own cloud controller?  Something else?  Not clear to me, but it's an important question as it's this critical piece of software that makes a public cloud a public cloud.

There's also the thorny problem of existing VMware EULAs.  VMware has sold a boatload of these.  Can these customers take their licenses to IBM's cloud, or do they have to buy again?  Details, details.

And, let's keep in mind, there are certain demanding workloads that many IT pros won't run on vSphere -- no matter how much powerpoint you put in front of them.

Caveats Abound

Proof-in-the-puddingAs they say, the proof is in the pudding.  We'll have to wait a while before we see an actual cloud service in the marketplace.  

What's not clear is how long.  A year?  Maybe more?  The market is moving fast.  My view is that -- unless you already have a viable enterprise cloud service in the marketplace today -- it's probably too late to start a new one.

And, not to be snarky, VMware customers have every right to be skeptical about yet another forthcoming VMware cloud strategy.  I've sort of lost count.  No, wait, this is the one ...

Finally, there's not going to be a lot of air cover from the mothership on this one: either EMC or (soon) Dell.  

Yes, VMware is independent, but when your corporate parents take a dim view of things, it's never easier.

Not to imagine some potentially awkward conversations. Customer to VMware/EMC/Dell rep "what's your strategy for hybrid cloud?"  Errr, call IBM?

The Inevitable Oracle Comparison

Oracle_cloud2At a high level, there are three fundamental components to Oracle's cloud strategy

  • A fully-featured and successful public cloud offering: SaaS, PaaS and IaaS -- plus a boatload more.  It would take you a week to go through everything that's there.
  • Providing compatible public cloud equivalents to what we sell on-premises, e.g. Exadata Cloud Service.  Think of this as "cloud insurance" if you will.
  • The ability to bring a fully-featured public cloud model on-premises.

All from a single vendor engineering the stack, and supporting everything.  None require a "strategic partnership" to be successful; Oracle controls its own destiny -- which I find somewhat reassuring.

What All This Means

It should be clear by now: cloud is the new enterprise architecture.  Public and private clouds need to work closely together to deliver the agile IT framework demanded by the digital economy.  

If you're a traditional IT vendor that can't show their customers how to easily and confidently get to integrated public/private, life is going to get much harder for you -- and your customers.

The future is coming fast.


Like this post?  Why not subscribe via email?




When A Patch Crisis Isn't

RecallIn the automotive world, a defective part from a supplier can result in expense and tragedy.  Witness the massive Takata airbag recall, affecting dozens of manufacturers and tens of millions of vehicles on the road today, including potentially yours :(

The IT world is no different, we are all dependent on components from others, especially Linux and open source code.  Bugs are found, some are serious -- and they must be quickly patched at considerable effort and expense, otherwise tragedy may await.

Last week, a particularly nasty bug was found in the widely used glibc code that enabled bad guys to essentially take over a DNS server.  More details here, here and here

The severity of the bug resulted in a "PATCH NOW!" directive to the IT community at large.  While not as nasty as the infamous Heartbleed or Venom bugs, this one merited a serious and immediate response.

For many IT shops, this sort of all-too-common fire drill involves not only a lot of effort, but downtime as well. 

Except for Oracle shops, that is. 

The Magic Of Ksplice

Changing_enginesTypically, it's very hard to patch code when it's running -- sort of like swapping out an airline engine while it's at cruising altitude.  Much easier to just land the plane.

One approach is to simply update the code on storage, and restart everything.  Downtime is never ideal, and you silently cross your fingers when things are coming back up. 

Unfortunately, that's so often the norm.

Another popular approach is to use clustering technology for fast failovers.  Update shared code on storage, and then failover each server in sequence using a clustering technology to minimize production impact.   Better, but it's still a process that deserves close supervision.

These days, there's a third approach, and that's Ksplice.

Ksplice is a Linux framework, pioneered by Oracle, that enables hot-swapping of code modules on-the-fly with ZERO disruption.  Old code is unhooked and new code is re-hooked -- and the application is completely unaware that anything has happened.

HookKsplice has been happily performing non-disruptive online patching on kernel code in Oracle Linux for several years.  It works, and it works great. 

But what about user space code, like the aforementioned glibc?

User space code is a much harder, as Oracle has to develop individual harnesses for specific libraries that are widely in use.  It's not a generic capability.  If you're running user code that has been modified with one of our Ksplice harnesses, you're golden -- otherwise you have to patch like everyone else.

Fortunately, the code in question already had a Ksplice harness as part of our distribution.  Anyone running recent Oracle Linux code (and subscribing to the appropriate service) was able to non-disruptively patch the defective library code with ZERO downtime and ZERO impact to production.

It's an impressive feat, one that I'm sure was appreciated by more than a few Oracle customers.

More on Ksplice here, along with additional commentary here and here.

It should be pointed out, it's the same Linux that's used in our Oracle Engineered Systems: Exadata, Exalogic, Exalytics, Supercluster, Oracled Database Appliance, Big Data Appliance, Private Cloud Appliance, Zero Data Loss Recovery Appliance and so on.

The Value Of The Red Stack

RedstackOne of the things I didn't fully appreciate before coming to Oracle is that owning all the pieces in an IT stack lets you do some pretty impressive things.

Like automatically hot-patching a critical bug across a vast army of Linux-based systems with no downtime.

While I do see the historical appeal of rolling-your-own IT stacks, it's hard to argue with the ease and convenience of a completely engineered stack.  Including nifty features like this one.

Just curious: for all you non-Oracle Linux admins out there, how long did it take you to find, patch and restart every affected system?

Just curious :)


Modern Cloudwashing

When "cloud" became A Big Thing several years ago, we were all greatly amused by vendors who simply added a "cloud" moniker to familiar offerings from a previous era.

If cloud is the new shiny thing -- and, as a vendor, you're flat-footed in having new offerings -- why not simply rebrand the familiar as "cloud"?

Hence the term cloudwashing - painting a thin cloud veneer over what is most certainly not cloud.

ToTheCloudMicrosoft's infamous "To The Cloud!" campaign earned justifiable scorn, for example.

Vendors weren't the only ones doing a bit of cloudwashing.

Within many IT shops, the notion of a private cloud became quite popular -- a cool thing to do. An awful large number of ordinary virtualized server clusters got internally -- and aspirationally -- rebranded as  "private cloud".

Fast forward to 2016 -- and cloudwashing is still with us, but in a different form. Early versions of cloudwashing were responses to clear gaps between expectations and reality.

Modern cloudwashing is no different.

I would argue that, when it comes to enterprise IT cloud strategies, we're still cloudwashing ourselves: vendors and IT shops alike.

Modern Cloud Motivations

Early cloud motivations were mostly about saving money and not having to deal with traditional IT organizations. While those themes certainly haven't lost their appeal, I think there's a bigger motivation on the table today: abject fear.

Attitude-2-abject-fearWe've all seen how disruptive digital-born business models can be. We will inevitably see many more. That's particularly worrisome if you're at the helm of a large, successful company that *wasn't* born digital.

Pick your industry, it doesn't really matter.

These same large companies often have huge expense lines for traditional IT infrastructure: plumbing and plumbers. Maybe you can't be as nimble as a digital disruptor, but you certainly can learn to use their weaponry.

And so it has begun: board-level and CEO mandates to reposition enterprise IT as more cloud-like, fully realizing that cloud is the modern architectural and operational model of choice --- irregardless of public or private.

Time to get serious about this cloud stuff -- like it or not.

Foggy Thinking On Enterprise IT Cloud Strategy

So many folks I talk to only focus in on narrow parts of the cloud discussion vs. embracing it architecturally. They see individual pieces and parts, but not necessarily how it all comes together.

A little SaaS here, a little private cloud there, maybe a handful of experiments on the popular public IaaS clouds, a bit of glueware -- we're good, right?

No, you're not. Historians of IT evolution can point to many times when key architectural pieces went in haphazard, and had to be ripped, replaced and rebuilt at great expense and with considerable amounts of time lost.  

You don't want to be that guy.

Put differently, the difficult architectural choices you make today will be around for a long time, whether they were the right choices or not. I am adamant in encouraging a view where the pieces work together vs. separately.

Don't cloudwash yourself :)

A-La-Carte, or Integrated?

We expect our on-premises IT stacks of applications, databases and infrastructure to work reasonably well together. Why wouldn't we expect the same of SaaS (software as a service), PaaS (platform as a service) and IaaS (infrastructure as a service)?

Plumbing-pipe-bigI would assert that SaaS modules that are designed to work together is architecturally more powerful than standalone choices where you end up being the application integrator across functional pillars. Here, I believe Oracle has a strong advantage.

Imagine an application that told you which sales reps had the most profitable customers -- and could investigate HR records to determine potential cause and effect? Trying to do that with standalone functionality (financials, CRM, HR, ERP, etc.) would be quite difficult.

But it's a natural outcome of SaaS modules that are inherently built to share information and logic with each other.

I would also assert that PaaS functionality that is designed to work with your SaaS choices is architecturally more attractive than trying to mash up two worlds that have never really met. Modern enterprises want to build and improve on what they have, and quickly. That's the goal.

I would further assert that PaaS that can detect and exploit underlying IaaS functionality is extremely desirable. One popular PaaS offering promotes the fact that it is infrastructure agnostic. To me, that means it doesn't really run well anywhere.

Computer_systemHere's my point: in the on-premises world, we've learned to crave architectural elements that naturally work together and exploit each other's capabilities.

Why should cloud be any different?

So, hey, all you standalone Iaas, PaaS and SaaS vendors -- my take is that you're practicing modern cloudwashing. Your thingie without all the other thingies working with it won't be very useful.

Unless your customers are willing to do the heavy lifting on their own.

Big, Hairy Enterprise Applications

Largest-spider-in-the-worldSo many cloud propositions pick off the relatively easy workloads: test and dev, VDI, workload consolidation.

Those are interesting, but not where the real action is.

Line-of-business VPs don't get too worked up over that sort of stuff; they reserve their passion for the big, hairy applications that power their larger enterprises.

So, is your private cloud up to that? How about most public clouds?

We won't be able to fully embrace cloud architectures -- both public and private -- until we're comfortable with their ability to confidently run the stuff that can get you fired.

My, the room got awfully quiet all of the sudden :)

Public And Private Clouds Working Together

Many larger enterprises will want an architecture where they have choices as to where things run: behind the firewall, on a public cloud, or frequently a combination of both. Hard to argue otherwise. However, popular private and public cloud choices just aren't designed to work together.

And that's a big problem.

BuellerEvery time I see a familiar infrastructure vendor touting hybrid cloud, I ask myself a simple question -- can they offer a reasonable compatible enterprise-class public cloud consumption option to what they're selling on-premises today?

So many can't. VMware? EMC? Dell? HP? NetApp? Cisco?


To be fair, I will give both Microsoft and IBM partial points in this category. Oracle gets full points, though.

The rest of y'all are just practicing modern cloudwashing in my book. And leaving your customers high and dry.

Curbing Cloudwashing: Asking Hard Questions

If getting to an enterprise cloud architecture isn't just a theoretical exercise for your IT leadership team, I'd encourage you to start asking vendors some harder questions:

Cloudwash- Tell me how you deeply integrate SaaS, PaaS and IaaS?

- Can you deliver architectural equivalency between on-premises and public cloud solutions? And support both ends?

- Do you have customers who are comfortable running their most demanding workloads on your solutions?

Yes, I'm setting you up for an Oracle approach. Because, right now, we're the only folks who actually can answer these hard questions.

You'll just have to decide for yourself whether these are the right questions to ask :)


Like this post?  Why not subscribe via email?



Email subscriptions powered by FeedBlitz, LLC, 365 Boston Post Rd, Suite 123, Sudbury, MA 01776, USA.