If you’ve been following the discussion about how applications are now being written to be more “cloud native”, you might have seen some terminology that was strange or unusual. This is because the operational model around these newer types of applications makes a few assumptions which are very different than applications and infrastructure in the past. Let’s look at a couple of the concepts:
- Infrastructure as a Software: While there will still be underlying hardware for servers (some acting as storage) and networks, the core idea is that the way to interact with these systems is if they are built primarily from software, which is decoupled from the underlying hardware.
- Infrastructure as Code: Using software development techniques to define, test and manage the infrastructure software. This means that teams use structured automation frameworks (e.g. Chef, Puppet, Ansible, etc.) instead of random scripts, and the code is kept in versioned, software repositories.
- Avoid Patching: Instead of patching existing systems, the core believe is that new “versions” of a system should be deployed instead. This avoids the feature-patch-dependency creep.
- Abstract where Possible: Being able to abstract specific components within an architecture helps avoid lock-in to a specific element, and eventually leads to more programmatic ways to interact with more parts of the system.
Terminology to Learn
Durable – “survives failure of its parts” – This aligns to Abstract where Possible. Elements will fail and this is expected, but the overall system should be durable enough to survive either HW or SW failures.
Declarative – “user of the abstraction declares the desired state, and the service providing that abstraction attempts to maintain the abstraction in that desired state.” – This aligns to Infrastructure as Software and as Code. The owner of a system defines (in programmatic software) what it wants it to look like and how it should act. The system should be managed not to vary from this state or “promise”.
Immutable – “unchanging over time or unable to be changed” – This aligns to the idea of Avoiding the Patching elements of an applications, or the popular “Pets vs. Cattle” notion of managed resources.
Ephemeral – “lasting for a very short time.” – This aligns to both the idea of Infrastructure as Software and Abstracting where possible. Some elements are only needed for short periods of time, or will only remain as long as a dependent element is in use. This is partially why people are beginning to like using containers, as they are designed like processes, to bringing them up or down is simple and fast.
Remotable – “is an interface that allows IPC (Inter Process Communications) among other things. As you may know, all apps (most of the time) run on their own process and cannot directly interact with apps running in other processes directly. One method you can use to create an intercation between is by using a
IBinder allows communication between those “remote” objects.
We discussed many of these topics on a recent The Cloudcast podcast – http://www.thecloudcast.net/2015/09/the-cloudcast-213-what-is-immutable.html
Part of my world involves attending many, many technical events each year. Some of them are large vendor-sponsored shows, while others are small gatherings and meetups. For the most part, the lead speakers work for vendors and are highlighting some product or technology that they sell. It’s part of the model where companies with large R&D budgets develop technology, and products are purchased customers.
Over the past 12-18 months, the background of the speakers has begun to change. I’ve noticed that more speakers are coming from “customers” and they are either highlighting interesting business projects or software that they have developed.
It’s not unusual that customers would be on-stage with vendors at events, as vendors like to showcase “real life examples” of companies that use their product. But more times than not, these examples follow the standard IT model of deploying something new and saving some money for the business. What is unusual, or maybe I should say “becoming commonplace” are customer lead stories about actually transforming the business via technology. Complex problems that require new ways of thinking, and customers that have executive support to do things differently. Maybe they are using open-source software. Maybe they are using public cloud services. Maybe they went around IT and have integrated a SaaS offering into one of their commercial products. They are big problems, which require big ideals to solve them. The types of that would attract top-level engineers to be part of the solution.
Other customer stories talk about how they have solved a problem and are now contributing the associated software code back to an open source community. And this isn’t just NetFlix or Twitter, it’s banks and pharmaceutical companies and brick-and-mortar retailers and insurance companies and industrial giants. The types of companies that everyone likes to say are completely restricted from doing cool new things because of regulatory requirements.
So what does this mean? Why has this trend been happening for the last 12-18 at events around the world? Continued »
It’s that time of year again, when all the little ghosts and goblins come running up the yard and ringing the doorbell, looking for treats. Some costumes are obvious, while others require a little bit of explanation. Either way, they all make you think you need to give them some of your candy. Right there, in your face, wanting to know just how much of your candy bowl they’ll be able to take from you.
This got me thinking – what would some of the Enterprise IT vendors be dressing up as for Halloween this year? Let’s take a look:
Oracle – Iron Man
This one is pretty easy. They are extremely well funded and have a strong set of powers. They obviously love being awash in red, with a glow of power around them. For many years, you hadn’t heard that much about them, but their recent cloud launches, just like the recent Iron Man movie franchise, have them in many people’s watch list.
They believe that you should give them most of your candy bowl, and they expect you to have at least that much saved for them in each of the coming years.
IBM – Donald Trump
Always impeccably dressed and well spoken, IBM lets you know that their costume cost at least $1B, because they don’t engage in any public activities unless that threshold has been committed. Like mainframes, Trump’s core business has been in brick and mortar real-estate and other lasting assets. But sometimes they like to make a splash, by partnering with companies like Twitter or Apple, to create a new impression in existing markets – sort of like “The Apprentice”. America isn’t quite sure if they believe in Trump, as polls have been down lately – sort of like IBM’s stock price.
Hewlett Packard (HP) – Centaur
Recently split between HP Enterprise and HP Inc. (or should that be ‘HP Ink’?), they are still somewhat connected by the common HP name. HPE wants to be blazing new trails, firing arrows into new markets. HPI is a steady business, moving forward like a plow horse dragging lines through a field to plant crops in the springtime. Do they co-exist, or so their separate ways. And can a centaur exist independently, without the other half?
Cisco – Batman
Wealthy and powerful like IronMan/Oracle, but often takes on the mild mannered appearance of their leader (John Chambers or Chuck Robbins). And like the recently plethora of movies made about Batman, Cisco has attempted to show the world different views of their existence – networking company, collaboration company, consumer company, Internet of Everything company. They’ll be watching your candy bowl, whether you know it or not.
The AWS re:Invent 2015 show was a couple weeks ago and I’m finally starting to get my thoughts together on the announcements and trends – got a little distracted by the Dell/EMC merger announcement last week.
Patterns are Emerging
While I always try and make predictions prior to the show about potential new services, I’ve learned my lesson and will start taking a different approach to thinking about AWS. Just like AWS has a tremendous system to collect feedback about their services from the end-user usage patterns, the rest of us should be able to start seeing certain patterns emerging. The number #1 pattern to recognize is “make it simpler, eliminate friction”. This is now obvious with services like Lambda, Elastic Container Service, Kinesis Firehose and the IoT service. While the uber-techies could explain how someone could use containers on AWS before, it was complicated. Someone could explain how to stream data from sensors to AWS, but it was complicated. The new patterns is to eliminate that complexity and allow it to be simpler to use by the masses. Services like Lambda answer the question of, “why do I need to manage an entire server or VM just to run a few functions?’ Services like Kinesis Firehose answer the question of, “how do I size the input and processing of my streaming data from IoT end-points?”
While some people will focus on things like Snowball for data transfer, which by-the-way is creative way to think about getting around the speed of light (“data gravity”), the foundational infrastructure pieces are all in place now. Time to get ready for the onslaught of combined services that will simplify how to build upon them.
The Enterprise Believes in AWS and Looks towards the Long-Term Horizon
Companies like Capital One and GE and Accenture and FINRA, who work in highly-regulated industries and have been around for decades are trusting AWS to build their next-generation platforms. In some cases they will use native AWS services as the building blocks. In other cases, they will use software platform layers like Cloud Foundry on top of the infinite AWS infrastructure resources. They have done their due-diligence on the platform, the security, the performance and are now voting with their wallets for frictionless access to resources and advanced services. These types of companies are the lighthouses for other large enterprises, and you can be sure that other CIOs will be making calls to their colleagues to learn from their successes and failure (just like at DOES)
“Hybrid Cloud” is being ReDefined
Every traditional vendor is betting their future on Hybrid Cloud. Their vision of the Hybrid Cloud is initially a refresh of their customer’s existing hardware to build a Private Cloud. And then add a bunch of Cloud Management Software that automate repetitive tasks and eventually add a self-service catalog of services that are available to internal developers. Eventually, those customers will hopefully be able to turn up some resources and services in a public cloud environment, but that might be constrained to only those public clouds that use the exact same technology – assuming any of those public clouds still exist in a few years. The implementations will be slightly different from vendor to vendor, but they are all following the same playbook. Continued »
The technology industry is at an interesting crossroads right now. At one end of the spectrum are companies like AWS that are growing incredibly fast and are reshaping some of the fundamental rules of the industry. At another point on the spectrum are the 100s of open-source projects and start-ups that are driving innovation and new approaches to both infrastructure and applications. The projects are also disruptive, but many people aren’t quite sure how (or if) they will become monetizable. Then there’s the stuff that is sometimes called “slow tech”; the large technology vendors and their customers that have dominated industry verticals for decades. These companies generate massive profits, employ millions of people, and generally feel threatened by many of the newer or faster technologies on the spectrum.
Fear creates some interesting behavior. It’s a nature element of survival, and people can do incredible things when their intrinsic survival techniques take over. I believe psychologists call it our “lizard brain”.
In San Francisco this week, the 2nd annual DevOps Enterprise Summit (DOES) was held and about 1200 people attended. These people were from 400-500 companies, coming to share their stories about making the dinosaurs dance and to share tribal knowledge with those that wanted to survive and grow.
For all the news we hear out of Silicon Valley about unicorns, DOES is a pragmatic event. It’s focus is on real businesses making real changes that create real financial impact. And these changes didn’t happen overnight. In fact, most of them took several years and are still a work in progress. No greenfield environments here. It’s a jugglers conference, with mainframe apps, legacy apps and cloud native apps acting as the flaming torches. Continued »
It’s been an interesting two weeks, hasn’t it. I was at the AWS re:Invent 2015 event in Las Vegas last week and things were gong along as expected. There were 19,000 people in attendance and AWS continued to announce new features and highlight interesting customer use-cases. To say that AWS is impressive is an understatement. They are powerful and professional and they are doing things that no other technology company has done to impact Enterprise IT in a very long time.
And about mid-week, there was a strange rumor that started getting momentum and cross-referencing:
“Dell to acquire EMC”
Huh? Dell, the company with all the debt from going private two years ago? What about EMC buying VMware, or vice-versa? What about the EMC Federation? What the Dell?
And then on Monday, it happened. Dell acquired EMC. It’s not officially closed – that’ll probably be next spring or summer, but October 12, 2015 is now official “Tech A.D.” – after Dell.
Sitting in the middle of these two events is somewhat disorienting. I’ve run out of fingers and toes for how many people called and said, “Did that really just happen? I’m having a hard time processing this.”
I’ve actually spent quite a bit of time this week processing, writing about and discussing this Dell/EMC merger (here, here, here, and here). There more I think about, the more I realize that there is potentially a new angle to write or discuss everyday for quite a while. But instead of talking about portfolio strategy or tracking stocks or branding changes, I thought I’d throw out the concept of Tech A.D. (credit Amy Lewis, famous on Twitter @commsninja, for the clever name).
When I was a guest on Speaking In Tech this week, Greg Knierieman (@knierieman) asked if the AWS re:Invent show was altered when the Dell/EMC rumors began. Stu Miniman (@stu) said, “Other than the press, nobody in attendance really cared.”
Let that sink in a little bit. At the fastest growing event in our industry, the vast majority of the attendees DID NOT CARE about the LARGEST IT MERGER IN HISTORY.
This is what inflection points look like. You may or may not care about Dell/EMC, but the ripples from this intersection of AWS growth and two IT titans merging will be felt for quite a while and will very likely have an impact on your tech career for the next decade.
Simple question – which IT model are you aligning yourself to for the next decade? The software-driven, OPEX-based model, or the solution-driven, mostly CAPEX-based model? Which one are you betting your company’s future, and which one are you betting your career?
At AWS re:Invent, you walk away thinking that there are some incredible new applications and businesses being enabled by this new model of computing.
After the Dell/EMC merger announcement, you walk away thinking this could be a trainwreck or a bellweather for change across the “traditional IT” segment of the technology industry.
The latter of those walkaways has lots of potential twists and turns. The former looks like a freight-train carry gold, but also running over many things that aren’t quite sure what to do on or near the tracks.
It’s going to be an interesting next couple of years. So what decisions are you making after Tech A.D.?
There are four big infrastructure events each year – VMworld, OpenStack Summit, Cisco Live and AWS re:Invent. While it’s not the largest in terms of attendance, in my opinion, AWS re:Invent is the most important because it is setting the tone for where infrastructure is evolving – both in terms of technology and economics. Heading into next week’s event, these are my thoughts about the (projected) AWS announcements and focus areas.
- AWS is focused on killing legacy IT and legacy vendors. Won’t focus on any of the other cloud providers (Google, Azure)
- Continued focus on companies that are “All In” about using AWS. This will be ISVs, and many verticals. The verticals won’t actually be “All In”, but segments of their business that are rapidly changing will be shown as moving all applications to AWS (Media, Healthcare/Pharma/R&D, Financial Services, Gov’t)
- They see weakness from the major vendors (EMC, NetApp, HP) and will announce big price cuts to EBS, S3, Glacier and Elastic File.
- Some focus on how more of the EC2 (VM) instances are completely backed by SSDs and how that pricing is coming down sharply.
- They’ll simplify how snapshots/backups for DB services across regions, further eliminating the value of storage admins. Saw this already with Aurora DB.
- Lots of focus on Elastic Container Services and Lambda. Trying to establish themselves as the next-gen cloud that is beyond VMs and now focused on micro-transaction for new applications
- Lots of focus on the growth of these areas and trying to highlight that Cloud Native apps are becoming more critical to business survival and success.
- There will be new focus on simplifying pricing, including ways to pre-purchase resources.
- They will have a focus on Enterprise Identity Services (e.g. Active Directory, LDAP) integration
- They will expand upon their Service Catalog offering, which is one of the areas that most vendor-led Private and Hybrid Cloud offerings don’t get right and most IT organizations don’t know how to create (they can’t make IT into “products”)
- Don’t expect much from them here. PaaS market is still fragmented and AWS doesn’t want to let Cloud Foundry define the model. AWS has always believed in the UNIX approach of offering multiple loosely-coupled or individual tools to developers and operations teams.
- Big focus on AWS Lambda (Micro-Compute) + Kinesis (Stream-Processing). Consumer, Manufacturing, Gov’t (Region/City) and Healthcare are super-focused on this space. Lines of business driven, not IT.
- Big focus on making mobile apps simpler to build and run.
- Focus on cross-platform application services, not just Apple or Android
To paraphrase the Dos Equis Most Interesting Man in the World, “I don’t always write about development platforms, but when I do, I tend to dig in fairly deep“. These new platforms are generating large amounts of interest and community following – as well as attracting large amount of VC funding.
An interesting element of that research is that very few of those platforms are coming from companies that are well-known in the application development space. But the market is moving quickly around this concept of “Cloud Native Apps” and almost every major IT vendor and start-up is trying to find a way to be relevant in this emerging space. As they all know, applications are becoming central to creating new business models (see: Uber, AirBnB, Netflix, Square, Facebook, etc.).
I recently predicted that Cisco might acquire Salesforce, but until that happens, they appear to be quietly building an application development stack. The platform appears to have two pieces:  Cisco Shipped and  Mantl. Both of them are open-source and are based on a variety of popular infrastructure and application tools.
“Shipped” is focused on the developer experience; getting applications from their laptop and moving them into environments for testing or production. It is focused on applications running in containerized environments and works with basic scripts and GitHub collaboration.
“Mantl” is focused on the complex backend-integration that’s required to run these Cloud Native applications (built on microservice architectures) across any cloud platform. It’s built using a broad set of open-source software frameworks and tools, many supported commercially by companies such as CoreOS, Docker, Hashicorp, Elastic, Mesosphere, NGINX, and Red Hat. Continued »
- Watching the beginning of College Gameday at 9am and seeing the pre-tailgate crowds piling up on campuses across the country.
- Listening to this jam.
btw – I’m still not sure that I need a little bang in my ting-tang….but I digress…
Saturday mornings have one challenge – what to do between 9:30am and 11:55am, when Lee Corso’s predictions are made? Usually the right answer would be to sit in front of a huge grill and make sure the temps stay around 225* and the smoke on pork shoulders is thick. But this morning we’re going to try something different – I’m going to make my own predictions. So here’s a few crazy ones certain to be wrong…
Cisco Buys Salesforce
Cisco has a new CEO (Chuck Robbins), who seems to much more focused on software than his predecessor. Cisco has made a bold prediction that the Internet of Things (or “Everything”) will drive a $19T opportunity over the next decade. To make this happen, Cisco needs not only a big presence in the Cloud, but a big presence in software applications. At Dreamforce this week, the Salesforce IoT strategy was front and center. How quickly with the Cisco board allow the new CEO to have access to the big checkbook?
Microsoft Buys Docker (and maybe Mesosphere)
Windows is the dominant PC operating system. But PC sales are on the decline, being replaced by smartphones, tablets and alternative server and laptop OS’s. Docker is becoming a favorite of developers on Linux, and Windows support is coming soon. And Docker has an astronomical valuation for a company that has shown limited revenues (to-date). Microsoft has the available cash and needs a way to attract developers in either Linux or Windows environments. Can Docker be the technology that helps Microsoft move forward from being OS-dependent in the future?
Oracle Buys EMC
Everyone is expecting some shakeup in the current EMC Federation model, with the heavy betting on EMC buying the rest of VMware and officially bringing them under a single umbrella – maybe even adopting the “VMware” name as the overall brand. That might happen and it would be interesting to see who would run that combined entity. But there’s an outside shot to consider – Oracle now has former EMC leader Dave Donatelli running their Exa-business. Donatelli wanted to run EMC at one point in time. Would the EMC Federation assets help Oracle broaden their markets, as well as taking out a fierce competitor?
When Amazon Web Services (AWS) is reporting $6-8B/year a revenues, it’s beginning to threaten some existing IT leaders. Just like College Football is beginning to get more comfortable with a more rapid pace of play, the world of IT is going to need to get used to faster changes going forward.
Like many people, I watched as my Twitter stream exploded yesterday with reactions about the latest announcements from Apple. New watch bands, faster iPhones, bigger iPads, updated AppleTV boxes and a magical pencil. Mixed in the announcements were some discussions about a few of the software features that could be used on these new devices.
A few weeks ago, I wrote a brief review of an excellent new book by Stephen O’Grady, called The Software Paradox. It highlights the shifts in how software is distributed, used and ultimately monetized (or not) by both vendors and providers of software-centric services. One of the core premises of the book is that it’s becoming increasingly difficult to monetize packaged/embedded software being sold directly to customers.
With that in mind, let’s get back to all these Apple announcements. What has Apple become? It can be argued that it is now a massive hardware company that gives away the embedded software for free. iOS or MacOS do not run on anything except Apple hardware, and increasingly, that software is now available on an on-going basis (e.g. updates/upgrades) for free. In terms of online services, Apple makes the bulk of their revenues by being the orchestrator of a marketplace (e.g. AppStore or iTunes), not by selling their services or their software. They take a 20-30% cut for running the marketplace; essentially a reseller of digital goods.
As much as people might want to mock the lack of “new feature innovation” coming from Apple, what they are putting together is a fairly brilliant strategy. They seem to realize that revenues from stand-alone software are difficult to create and grow. They understand how to create experiences when hardware is bundled with software, regardless of who builds the software. They’ve created a user-base that is reasonably locked into the Apple experience and has been mostly conditioned to update the required hardware every 2-3 years. And since Tim Cook is a genius at supply-chain-management (getting the silicon turned into products), Apple is able to exert massive control over the pricing of the underlying hardware elements (or “COGS” – Cost of Goods Sold).
The parallel to this strategy in the Enterprise IT market is the growing trend to bundle together hardware and software under categories like “converged infrastructure” (CI) and “hyper-converged infrastructure” (HCI). While open-source software is often out innovating commercial software, it still struggles to make revenues when sold directly as a commercial product. But bundle that software together with a defined set of hardware and a known end-user experience and you can begin to understand what products like VCE’s Vblock are at a $2B run-rate.
We’ve heard it ad-nauseum for the last few years – “software is eating the world“. And while it’s true that entire industries are being reshaped by innovative uses of software, it still requires the right packaging, the right experience and the right business model to make that equation successful. Sometimes we forget that simplicity usually wins, and those that take the most friction out of the “idea to execution” equation often get to keep the bulk of the value creation.