Skip to content

Why The Future of Business Is Reactive

showing a response to a stimulus.
“pupils are reactive to light”

A few months back Typesafe, one of our clients, called us to tell us about the latest hipster hacker term Reactive Programming, which is about new stacks designed for the new spiky, responsive workloads that so many of us are facing. My first thought was dismissive- “reactive? that’s bad, right?”

If you came into business any time in the last 20 years the mantra was always be proactive, not reactive. But does that really work any more.

But then I looked into what the geeks were saying about Reactive, and it began to make sense to me. After all – we’re now living in an age when so many of the traditional IT disciplines are falling apart – capacity planning, for example. Everything today seems increasingly bursty. People expect real-time now.

I started wondering what a reactive database would look like – it would be schema-less, for one thing, so completely flexible from a data model perspective.

Typesafe uses Reactive as a description of its stack.

“The Typesafe Platform is a JVM-based runtime and toolset for building Reactive Applications”, where Reactive is defined thus:

Reactive Applications are a new class of applications that are becoming more and more prevalent in both Consumer and Enterprise-facing environments. Reactive Applications are fundamentally different to the traditional web-based or mobile applications seen today and are distinguished by having one or more of the following defining traits:

Resilient: The ability to recover and repair itself automatically in order to provide seamless business continuity.

Interactive: Rich, engaging, single page user interfaces that provide instant feedback based on user interactions and other stimuli.

Scalable: Can scale within and across nodes elastically to provide compute power on-demand when it’s needed.

Event-Driven: Enables parallel, asynchronous processing of messages or events with ease.

the TypeSafe framework is used in high frequency trading and spike-oriented architectures for next gen media companies – with apps built in Scala, a functional programming language that runs on the JVM. But this post isn’t about that. It’s about the core idea.

Because the more I thought about Reactive the clearer it became that businesses, not just infrastructures, need to act in this way. The age of long term planning is long gone. Circumstances change and they fast.

Google didn’t set out to revolutionise the advertising business.

Blackberry didn’t expect to lose more value in a year than any company in corporate history. But then, Palm used to be the game too.

According to Mark Andreessen Software is Eating The World – all industries are going to be remade in the way Google changed advertising. Even IBM is now warning its own clients that their competition won’t come from existing players, but from new Web companies. The nature of competition is changing.

I think Stowe Boyd describes the new new thing better than I ever could:

The takeaway is we’re now in a time of overwhelming volatility, uncertainty, complexity, and ambiguity, where forecasting becomes nearly impossible, and even determining which of the risks confronting us are most significant is almost unimaginable. This near blindness confronts governments, global businesses, the local bank, the entrepreneur, and every person on Earth. We are driving a car whose engineering is unknown (and rapidly evolving, in real time, under the hood), a car that is inexorably speeding up, with only one flickering headlight, and no brakes, on a road headed to who knows where.

Businesses need to be reactive because you can’t predict the future and they will need new technical architectures to support the change, which look a lot more like Web computing. Agile, bursty, lean – that’s the future of business.

In the mean time, check out the Reactive Manifesto

Categories: Uncategorized.

Opinionated Infrastructure: Hybrid Cloud Made of Prius and Zipvan

In which I riff on the cloud again. Pay as you go? Radical simplicity or interim solutions? Turns out the Toyota Prius can tell us something very interesting about the value of cloud, and whether we really just all move to the future, or whether we need to evolve infrastructures towards it. To whit – inefficiency isn’t always bad… some cloud heft may not be a bad thing. Data centres are not all bad.

My client IBM PureSystems, which sponsors the Opinionated Infrastructure series, can now span public and private, post IBM’s SoftLayer acquisition, which is an actual public cloud.

Categories: Uncategorized.

On Confirmation Bias

too good not to share, here.

Categories: Uncategorized.

A Moderated Screencast: Enterprise Cobol for z/OS 5.1

IBM mainframes have always been great at backwards compatibility, but the compatibility comes at a price. While an app developed many years ago will generally run faster on the very latest systems hardware and software, it won’t be able to take advantage of the latest features- such as XML or Java acceleration through algorithms or specialist processors. With a major refresh of its compiler however IBM has addressed this issue, so that now mainframe applications can offer what I like to call forwards compatibility, often through a straightforward recompile. Cobol just got more useful – supporting modern data structures and bringing significant performance improvements. In mainframes many customers pay to the peak load, rather than the average, so anything that shaves those peaks can save a lot of money.

Hope you enjoy this show, sponsored by IBM.

Categories: IBM.

Opinionated Infrastructure: Searching for Hadoop, The Money Counter Episode

Hadoop is an extremely powerful platform for certain data management tasks, counting and sorting huge amounts of data, notably from unstructured data sources. The entire database and analytics industry is currently retooling for Hadoop, but that means customers have a lot of choices on their plate. In this video I explore the idea of counting and sorting as a service, and related packaging implications. In gonzo style of course. Whether buying an appliance, a service, or rolling their own the business is going to have high expectations about Hadoop service levels and capabilities. The trick is to balance risk against opportunity – and make the right package decisions. In any tech wave the best packager wins, and wins big, because they make things convenient for users.

disclosure: IBM sponsors this video series, but editorial responsibility is ALL MY OWN.

Categories: data.

Tags: , , ,

Opinionated Infrastructure Refreshing The City: Converged Infrastructure and Legacy

The regeneration of Shoreditch can tell us something about systems architecture and the benefits of legacy. What is modernisation? Thoughts on governance, trains, networks, and pace layering of legacy. Inspired by the converged infrastructure my sponsor IBM PureSystems is building.

You should watch this video.

Categories: IBM.

What if IBM has a clue: Get Lucky

It’s easy to knock IBM, and a lot of developer types do. But IBM was not born in the era of digital abundance-driven developer primacy, rather from a time of top down purchasing of expensive machinery in an age of resource constraints. One thing I have noticed in my 18 years of covering the firm however, is that the firm has an amazing capacity to Get Lucky. Key competitors seem to stumble and then fall over at key moments.

Companies that should have hurt IBM include HP, which acquired a dominant position in a market IBM chose to abandon – PCs – and then lost its focus on commodity gear as it did things like buy the mainframe bad bank otherwise known as EDS.

Or Veritas, which was killing IBM until Symantec acquired it.

Or Dell, which has spent the last 10 years trying to escape the commodity trap – a trap that IBM never falls into.

Or VMware, which is currently in a weird holding position.

Or Oracle, which keeps buying companies, but keeps seeing software revenues decline, and equities analysts discount its shares.

Or Sun, well you know about Sun. It beat the crap out of IBM in the late 1990s then spectacularly self-immolated.

Or Microsoft, which had IBM on the back foot for a while, before losing focus.

But is this all really luck, or something else? As Louis Pasteur said: “Chance favours the prepared mind.”

Anyway some recent moves at IBM strike me as substantive, surprising, developer-focused and worth paying special attention to. A key driver behind these deals is Gerry Cuomo, an IBM middleware CTO recently given a business strategy role. Also keep an eye on Angel Diaz, who has set himself up as IBM’s mover and shaker on the West Coast. If it’s happening in the Valley, touches Cloud stuff, and IBM is involved, then it has Angel’s fingerprints on it.

So what has IBM done lately?

Adopt OpenStack

Acquire Softlayer

Adopt CloudFoundry (and quietly make a play to pwn it)

Adopted Mongo, to reach a developer audience it never could with DB2.

Start using Netflix tools like Chaos Monkey, adopting open source technology “given away” by what would normally be an IBM customer, not a tech provider.

Adopt Node.js – this is happening, IBM just hasn’t made a fuss of it. There will be a substantive play.

Persuade Google that POWER might have a role in high scale, low power data centres, bringing it into an alliance with IBM, NVidia and Mellanox. This last one is perhaps the most important point on the list. IBM can’t succeed in the enterprise if it can’t succeed at Web Scale. Is the target really Intel, or Facebook and its open data centre specification Open Compute? I would argue some of both. As Stephen has eloquently argued – hardware now faces some of same shifts as software. When Your Customer is Your Competitor: The Return of Roll Your Own.

As I often say, the best packager in any tech wave wins and wins big. Well IBM seems to be positioning itself as a potential master packager, helping enterprises to adopt technology born on the web, without worrying about IP and so on. Of course huge challenges face IBM- the small matter of Amazon and Google in cloud computing for example.

But In this business you make your own luck. If IBM is to succeed in the new era it needed a far better Developer Led Adoption story, and a better approach to embracing Web technology. Recent moves indicate it’s moving on both fronts.

Get Lucky.

disclosure: HP, IBM, 10Gen, Symantec, VMware are all clients.

Categories: Uncategorized.

Opinionated Infrastructure. Wheatgrass and Ginger: Thoughts on Hybrid Cloud. Lulz.

Here is the latest in my Opinionated Infrastructure video series, sponsored by IBM PureSystems. As ever, I got straight for the metaphor, and even by my standards, the shot shot is gonzo. Legacy isn’t bad, hybrid is good, and design is everything in making different modes of cloud work together. The location? Same as Luther of course… Shoreditch.
If you’ve been following, the way this works is to riff on stuff for Opinionated Infrastructure, which then flags my live PureChat series of GoogleHangouts, where industry luminaries talk to the trends. The next PureChat is September 12, at 5pm EST, and we’ll be looking at Hadoop.

Categories: Uncategorized.

Do we need to radically simplify infrastructure in order to do DevOps goodness?

I posted the other day about the amazing speaker line-up on my recent PureChat, sponsored by IBM PureSystems. But seriously folks- the content rocks. These folks really know what they’re talking about.

I love the stuff from Netflix’ Adrian Cockcroft, where he makes clear that getting the organisation structure right is more important than underlying infrastructure if you want to drive Continuous Deployment, supported by DevOps.

Check out Luke Kanies, Puppet’s dad, talking up production Infrastructure. He doesn’t like the term legacy. Automate what you have!

15 mins of goodness. You should watch this video.

Categories: Uncategorized.


MOAR VIDEO: Heavy hitters in DevOps and Continuous Deployment

We recently recorded this episode of PureChat, my IBM PureSystems sponsored Google Hangout, and the folks, and the guests are just stellar. I mean absurdly so. One interesting thing – pretty much everyone on the call is using Netflix tools like Chaos Monkey in their operations. Netflix is one of those new kinds of companies Stephen has been talking about – the value is not in the software folks, it’s in the service and the data.

  • Adrian Cockcroft (@adrianco), Cloud Architect at Netflix
  • Dave McCrory (@mccrory) SVP of Engineering at Warner Music Group
  • Luke Kanies (@puppetmasterd), Puppet Labs Founder & CEO
  • Kohsuke Kawaguchi (@kohsukekawa), CI Jenkins Server Founder, works at CloudBees.
  • Adam Jacob (@adamhjk), co-founder of Opscode and creator of Chef
  • Dan Berg (@dancberg), IBM DevOps Chief Architect
  • Robbie Minshall, IBM Rational Cloud Architect

IBM cut the video into chunks, so if you prefer smaller pieces, loosely joined, check out this post. Tariq Ahmad from IBM also helpfully wrote out the quotes below for his post. What is more.

“When we (Netflix) did the Cloud move, think of it as looking at every step in the process of getting value to customers, from having an idea to deploying it.” — Adrian Cockcroft

“There are a lot of companies, including Warner, that envision all of this coming together and transforming their business. The difficulty is in aligning the business needs with how you can automate.” – Dave McCrory

“I started Puppet Labs because I looked around and said oh my God the world of operations hasn’t changed in 10 to 15 years, and I don’t see anything that’s going to make it change any time soon. Let’s build tools that help people get rid of old technology and replace it with new technology.” — Luke Kanies

“We really haven’t figured out how to share what people do in one context to another. We have these conferences where people come in and talk about the amazing things they have done. But for someone else to replicate that work, it’s basically all custom made from scratch.” — Kohsuke Kawaguchi

“The Netflixes and the Amazons of the world and all the rest of what’s great about the Internet, sort of showed us a different way to deliver value to customers. We were doing it on the internet, we were doing it on mobile devices. That value is being delivered digitally rather than by going to a bank and talking to a teller.” — Adam Jacob, co-founder of Opscode and creator of Chef, speaking on legacy systems

“We don’t have one, two applications. We have thousands. How do we manage those? How do we increase our ability to deliver more rapidly? How do we deal with the complexity of our applications? Dan Berg

“There’s certainly a lot of value that folks find in certain standards, especially when we’re talking about platforms and applications that run on a variety of platforms. But there’s also a space for innnovation, to the side and also on top of standards that we think in the competitive space that’s really important to us as well.” — Robbie Minshall

Categories: Uncategorized.