Technology

Open Sourcing Is Human Nature by James Kelly

Yesterday was Earth Day. It’s a time when we awe at our world, the good and the bad. I get the feeling that with every passing year Earth Day grows in importance. Maybe it is all those Netflix documentaries I've consumed starting from “An Inconvenient Truth” and on down the list, maybe it is the yogi in me, or maybe it is just plain to see that Mother Nature is wounded and begging us to look beyond the economy to the ecology of our habitat, our impact on it, and each other. Dr. David Suzuki says, our collective egos get caught up in progress for progress’s sake, creating false externalities (boundaries) on our impacts, instead of remembering that everything is connected (I had a another beautiful reminder of this when I recently re-watched Avatar).

Instead of thinking about changing lightbulbs this week or thinking about what to cut out, I asked myself, “How I can better the world using my own talents and passions?” Earth Day seems like the perfect opportunity to reflect.

Because I've been thinking about an article on technology to kick-off my revamped blog here, when yesterday by odd coincidence an idea sparked in an unlikely reflection on nature, I jumped on it...

Technology and nature? Yes! A software-focused technologist by vocation, yesterday, my above train of thought collided with my thinking on open source…

For those friends that need the background… Defn Open Source: In production and development, open source as a development model promotes universal access via a free license to a product's design or blueprint, and universal redistribution of that design or blueprint, including subsequent improvements to it by anyone.

Yesterday, I was reviewing the newly published 2015 Future of Open Source Study. (Even if you’re not in high-tech like I am, I bet you have heard of open source software like Android, Linux/Unix, and maybe WordPress, SugarCRM and OpenStack to name-drop some prominent examples and of late Microsoft is getting into it too.) I was reviewing the study when I began to ponder how the growing momentum in the open source model that I know so well for software like the above examples can be applied to answer my question in general… how the future of open source (and my advocacy for it) is important for the future of the Earth and our lives here. Could open source even be adopted beyond the realm of software?

I didn't have to consider that for long before I realized…. Of course! Open source has applications beyond software. Humans have long been openly sharing things like education, cooking recipes, stories and books in libraries, and much more (hello! the wheel). However, more importantly I realized that open source is in our nature, aligning with a recurring pattern of successful species and group behavior and decision making…Birds flock together, fish swim in schools together, animals pack together. In all cases science has shown that group consensus decision making and a democratic model is present, and it leads to the success of a species and the good of the whole. This wisdom through successful trials has gotten baked into our DNA.

Have we lost touch with our nature?

So why isn’t open source always the default model? Have we lost touch with our nature? I think so. I think just like we have wounded the Earth, not realizing it is our very mother, we are sometimes equally delusional in believing that the closed competitive structures we create are real.

Obviously there are the supporters of closed and proprietary source that point to creating value for “the shareholders” (almost as vague an excuse and unreal an entity as “the economy”). Also there are proponents in our unconsciously capitalistic culture that see open source tradeoffs through a different colored lens, the lens of business. Peter Levine an advocate and experienced open source software project founder himself, describes “the success or failure of open source is not the software itself, but in the underlying business model.” However, this view taints the way to look at the open source model in general, by tying and measuring it in economic impact and corporate winning.

Having moved south from Canada to Silicon Valley and America 5 years ago, I know too well the fast culture of capitalism, corporations, and its demand for consumerism and profit. A problem that is at the root of many other problems as well examined in "Capital in the 21st Century" (a must-read on inequality, capital, and democracy) and the easier-to-digest documentary “I Am” where Tom Shadyac asks, “What’s wrong (and right) with the world and what we can do about it?”

Summarizing the lessons of “I Am” and my own life learnings in democracy, holistic health, philosophy, yoga... one detriment to society at large is the idea that competition is of great value. This stems from the ego of course and the delusion of separateness, but it was further propagated by well-touting “survival of the fittest” when in fact Darwin’s book showed that to be of little value.

What is of value? What is right about the world? What is worth our awe? It turns out that care, compassion, and above all collaboration and cooperation are responsible for the success of species…including us, humans. We aren't the biggest, strongest, fastest… but we are highly cooperative. We are born because of communities and into communities. As Desmond Tutu put it, “We are, because we belong.”

Open source in service of cooperation is utterly ingrained in human kind’s most basic nature.

If we believe this and in a world where we and everything are all connected—The universe’s “Internet of EveryThing”—as most spiritual paths teach and even science sees in GCP and quantum physics, what can we conclude about open source? It seems obvious to me that it aligns with the wisdom proven out in the evolution of life. Open source in service of cooperation is utterly ingrained in human kind’s most basic nature.

In a culture of innovation built around competitive corporations, open sourcing at very least hedges us against our collective corporate egos. At best, it can introduce a community's sense of greater meaning, humanity, heart, love, and compassion into the business place and across them. I am reminded of playing hockey as a kid. My dad used to coach me and his teams to play for fun, not to win. Competition was (is) a game. It has a lot of benefits like a score to measure oneself against and improve, but in the end we're all out on the ice (on this Earth) to enjoy experiencing growth through different perspectives, or the duality of life. A nice corollary is that when we intended to have fun, we also improved and often won (did our best work).

Having used, developed and evangelized open source software, I always had an altruistic feeling that it was the right thing to do for many reasons. But now I see beyond just the practical and business reasons. In a world seeking and needing more consciousness, the open source model aligns with democracy, citizenship, and the natural order of things.

In software and high-tech at large, there has been rising strength in open source projects and communities. It is nearing a watershed moment as the preferred way to collaborate on and use technology, not just in software but also in hardware with projects like the Open Compute Project (OCP).

Let’s make future Earth Days and everyday day closer to a global consciousness, not simply a matter of LED bulbs and reducing, but also being conscious about what we are adding and choosing. For me that is a passion and advocacy for the open source movements. I encourage you to partake if you can, and if not, simply use them (vote with your wallet). There are all kinds of documented practical reasons, but now you have one more!

Over the bottom half of this decade we'll see open source change the technology industry, and I can't wait to see what other industries will tap into our nature and collaborate out in the open.

Bullish Momentum on Hybrid Cloud Bodes Well for OpenStack by James Kelly

As I hinted in my last blog, RightScale’s State of the Cloud 2015 Report uncovered that moving to cloud is now a given for pretty well all IT organizations. The extraordinary take away for me was that 82% of surveyed organizations say that their sights are set on hybrid cloud.

If you don’t understand WHY this is happening, have a look at my blog entry on hybrid cloud: what why and howwith plenty of cross-referenced articles.

This fundamentally bodes well for OpenStack, but we don’t yet see it at 82% adoption…why not? There is the obvious fud out there about OpenStack immaturity, but that is easily disproved by example. More importantly, I think there’s still a lot of cloud-washing, but even with respect to hybrid cloud. Is this hybrid cloud end state as simple as using multiple private and public connected clouds? No, but I assure you there is a lot of that going on, and those folks saying they’re hybridized. Getting to that point is great, but it is not the end game.

To realize use cases like split-tier apps, split-tier storage (shout out to Google for their nearline storage solution), multi-DC DA/DR for multiple apps and storage, best venue cloud moves and cloud bursting, we need to get closer homogeneity (-explanation-) at the IaaS layer, and probably an iPaaS for higher-in-the-stack PaaS and SaaS integrations. At the IaaS it’s a given that it’s going to be OpenStack, otherwise you’ll be in a small corner you’ll have painted yourself into.

If more folks needed to execute on these sophisticated use cases in short order, I predict the OpenStack adoption would spike. Alas, there are some IT orgs still fighting fires and keeping the lights on. As the wider IT “tide” rises to greater levels of automation, all the OpenStack “boats” will rise because of their amenability to the hybrid use cases (simple and sophisticated) on the broadest set of public clouds, providing the best outcomes in economics and choice. But this tide is rising slowly, and thus so is OpenStack. Things bode well for those notice it, less so for those that don’t.

Please Show Your Support

In this last month since I asked for your voting support, we’ve been overflowing with great news around OpenStack…

First and foremost thanks to our OpenStack Summit session proposal voters! Cutting to the chase…many session proposals were accepted... Remarkable success!

OpenStack Vancouver 2015 Sessions

  1. Building a Secure Multi-tenant Cloud for SaaS Applications: Challenges and Lessons Learned – We’ll be presenting with our customer friends and OpenContrail advocates from Symantec, WorkDay and Lithium (hey Lithium... thanks for powering this blog )
  2. Delivering an End-to-end Automated and Carrier-class NFV (Network Functions Virtualization) Use Case – We’ll be presenting with our customer friends and OpenContrail advocates from NTT-i3.
  3. Deploying a Hybrid Cloud with OpenStack: Dynamic Service Automation – We’ll be presenting with our technology alliance partner friends and OpenContrail advocates from Ubuntu, Supported by Canonical.
  4. Service Chaining Using Neutron Networks Implemented as Standard Compliant Layer-3 VPNs – Our OpenContrail advocates and customer friends at AT&T will be presenting.
  5. OpenStack Ousts vCenter for DevOps and Unites IT Silos at AVG Technologies – Our OpenContrail advocates and customer friends at tcp cloud will be presenting with their customer AVG.
  6. NFV – VNFs and Juniper's Contrail with Ubuntu OpenStack – We are pleased to join our alliance partner Canonical, Wednesday May 20, 2:40 - 3:20pm
  7. External sessionOpenContrail Users Group – We are please to continue our strong support of OpenContrail by sponsoring this OCUG event that will be held just down the street on Wednesday morning.

These are all MUST-SEE sessions, so hit the links and add them to your summit calendar straightaway.

More news on the summit

We are gearing up to support our booth with a few demos, and we’ll have a conference room available for private meetings. Should you want to chat, a private test drive of OpenContrail, or the full Juniper Contrail Cloud experience, let us know, and we’ll pencil you in for a meeting. Otherwise, be sure to stop by the booth. We’ll have some cool surprises and invites to be had, but more on that closer to the actual event.

Juniper In the news

Quite literally in the press we’ve had some great updates.

  • Canonical-Juniper partnership: We announced a strengthened partnership with Ubuntu- Supported by Canonical right around MWC2015. What’s been in the works as Contrail Cloud for some time (Juniper’s OpenStack + OpenContrail + Ceph + Server Management) is now backed by Ubuntu OpenStack and Ubuntu OS engineering and support proxied through Juniper. Furthermore, Canonical’s offering of Ubuntu OpenStack is thoroughly proven out with OpenContrail and Juniper Contrail Networking. More…
  • Mirantis-Juniper partnership: With a complementary webinar co-hosted by Mirantis and Juniper, last week Mirantis announced the completion of a documented and proven MOS + OpenContrail reference architecture. Catering to DIYs, Mirantis Fuel automated deployments, or Mirantis professional services deployments, Mirantis is embracing open networking with the leading open SDN solution based on their customers’ demand.More…
  • Orange Business Services (OBS) was in the SDN-MPLS World Congress news coverage by Light Reading, on record publicly saying that they’re going with Juniper Contrail Networking as their cloud networking and NFV solution of choice. 

This was originally published on the Juniper Networks blog.

From Private to Multi to Hybrid Cloud in the Enterprise - What, Why & How by James Kelly

From the results of recent RightScale and Gartner surveys on cloud, hybrid cloud is the ideal goal for about 75% of enterprises, especially with increasing importance on the near-seamless compatibility with the quickly growing public side. Let’s break down the differences between the private cloud, multi-cloud and hybrid cloud, so we can define them, investigate why hybrid is the destination, and see how to climb to such hybrid cloud heights.

 Private cloud: In this case we’re looking at a single in-house data center or site of data centers, orchestrated as one internal private cloud. This is a building block for hybrid and multi-cloud.

Although my blog title begins with private, private clouds are less common than you may think because IT shops will often puff up their chest so to speak, to inflate their cloud ego that their virtualized data center is a private cloud. Knowing that enterprises and vendors do “cloud wash” and mix virtualized and software-defined data center messaging with cloud messaging, we probably ought to define our cloud in less “cloudy” terms. The NIST definition of cloud computing offers a concrete definition that includes the characteristics of on-demand self-service, broad network access, multi-tenant resource pooling, rapid elasticity, and measured service (SDN Era Side Note: notice how the values behind these traits align with the values of SDN – I guess that’s a separate blog though). While cloud service and deployment models differ, traits do not. Put differently, one can think of the characteristics of prominent public clouds like AWS, and ask if one’s private data center IT meets those same principles. On the journey to really achieve these cloud characteristics, different enterprises are at different places and moving forward at different speeds.

I often liken the journey to climbing Mount Everest. There are a number of basecamp milestones that one usually passes en route to the peak. Most enterprises are taking these milestones as steps in turn from their history of being entrenched in the valleys of IT silos. Some have the luxury of little legacy IT and are able to leapfrog to higher basecamps, just like some climbers take a helicopter ride as far as possible up the mountain. In any case, the path is not simple and straightforward, and without proper preparation in future-proofing for the final destination, like climbing, what gets you part of the way, may not work to get to the summit.

 Multi-cloud: Multi-cloud is the notion of multiple interconnected private clouds. The clouds are interconnected with unified monitoring, management, and automation across globally distributed private clouds.

 Hybrid cloud: A lot of enterprises—at least those not constrained by internal data locality regulation—are already using public IaaS and PaaS cloud resource elasticity or virtual private clouds—managed or hosted by a public provider. To elevate their game to hybrid cloud, they need a better way to integrate management, or in other words hybridize it, with their multiple private clouds. Like the multi-cloud notion above, they seek to unify and orchestrate applications, software platforms, and infrastructure resources with a centralized system working across private and public clouds.

Why Hybrid?

The purely private, in-house multi-cloud models described above necessitate a reserve and capital expenditure beyond the current needs of the organization. Furthermore, private clouds don’t offer the public cloud’s as-a-service economics of choosing OpEx over CapEx. Beyond economics, the behemoth size and multitude of public cloud providers offers enterprises of any size a way to increase the resiliency and globalization of their IT footprint.

Personally, I related to hybrid cloud like my credit cards. They give me the elasticity to buy whatever I need on credit margin rather than saving. It can be a lifesaver in an emergency. For a businesses, not buying, but rather investing, elasticity is crucial to business agility to seize a business opportunity. As they say, “luck favors the prepared.” Fully realizing hybrid cloud, enterprise IT can prepare by design for infinite scale.

Getting to the summit: Realizing Hybrid

Getting to real hybrid cloud requires embracing the open source ecosystem. I realize that’s a bold statement. I don’t mean enterprises need to partake in open source projects, but they do need to heed them. Let me explain.

There is a saying I like going around that, “cloud is the new computer.” You’ll recall the terminal-mainframe model, then the client-server model, and now the app-cloud model. Cloud is the new application backend.

We know that COTS servers and storage is the assumed basis for cloud hardware. The network hardware is best served by providing a standard (ubiquitous like COTS) IP fabric, as long as it is scalable. Don’t read this as me saying there’s no innovation happening in cloud hardware. Quite the opposite is true, but the interface it provides is increasingly homogeneous.

The software side of cloud infrastructure is shaping up much the same way whereby a homogeneous platform will benefit applications (the ultimate driver of IT value) and devOps (drives application value faster). Remember the server software battle between Windows Server, UNIX variants and Linux? When Linux became the clear winner, it provided an obvious platform of choice for application developers. Much the same thing is happening today in cloud. If cloud is the new computer, ask yourself what will be the new cloud operating system. The only cloud management platform (aka OS) gaining momentum this year is OpenStack. Just passing its fourth birthday, it now has many robust distributions. I think there is a parallel to be drawn between the previous battle and this one for the cloud operating system. The OS with the largest open source following won the battle. The reason why open source wins this battle has many variables, and nowhere is it more important than for cloud infrastructure on which enterprises require application, application platform, and devops-automation portability across multiple private clouds and the public clouds. For enterprise cloud, the choice of an OpenStack-based cloud-OS foundation clearly aligns them for hybrid cloud success because by design, the OpenStack API hybridizes well with the leading public clouds, namely AWS, GAE/GCE, and increasingly Azure. Furthermore, OpenStack is increasingly the choice cloud OS of the other public, managed, and hosted cloud providers like Rackspace. To summarize what Randy Bias at CloudScaling covers in his Hybrid IT webinar, running OpenStack for the private cloud is the simplest path to hybrid cloud which, for the portability described above, requires approximate hybrid cloud parity between clouds in two ways with respect to a standard infrastructure and interface: functionally and non-functionally. Functional compatibility is required in APIs, in infrastructure (e.g. object storage, block storage, compute, VPC networking), and in other behavior (e.g. configuration semantics and default settings). Non-functional compatibility centers around availability, performance, and QoS. A third axis would also be around the economics because the public vs. private clouds must be roughly the same economically, so that neither is prohibitively expensive to use. OpenStack doesn’t guarantee these parities, but it is the best place to start because a number of distributions and integrating technologies (even from VMware) are focused on these parities, understanding that hybrid cloud is the best objective for customers, and that they drive sales.

I have been saying for a while that cloud infrastructure is simply too important and ubiquitous not to innovate in open source, with open standards and open interfaces, but I must say, I love how Jim Whitehurst at Red Hat draws an analogy for executives from the national interstate highway system to talk about standardization.

So back to my original statement, “hybrid cloud requires embracing the open source ecosystem.” Why “requires?” After all couldn’t we have a highway system that isn’t all freeways, but instead all express toll routes? I wouldn’t want to, but I suppose we could. They wouldn’t be as popular as they are, and they wouldn’t have revolutionized the logistics industry and our travel and commute accessibility. Like I indicated above, the cloud transformation for enterprise IT is done for the sake of a few things. Firstly, it’s motivated by devops automation, built out of programmability, best with open source, and whose portability is hampered by heterogeneous APIs. Secondly, it’s motivated by applications and application platforms. Many of these themselves are increasingly open source and designed for cloud-native scale out. These open source developers will choose to align with communities of their open source kin when it comes to choosing cloud OS APIs so as to optimize their applications for cloud. Furthermore, choosing enterprise cloud infrastructure is vastly more important than choosing something like, say, a smartphone. We know that open sourced Android is a great model for application portability across phone vendors and it has dwarfed Apple iOS in its worldwide popularity. Still, there are a lot of iPhones out there and some people have happily bought and locked into Apple’s ecosystem. Getting bought and locked into closed cloud infrastructure is prohibitive for an enterprise on a far more massive scale of investment, and it’s especially unwise if you’re betting against aligning the public cloud providers building their clouds on open source-based infrastructure.

Please leave your comments and thoughts below on what you think will rise to the top as the new cloud operating system. I know it is a journey of a thousand miles, and some enterprises are just taking the first step. Heck, some are still looking at the map as the technology landscape evolves! Whatever the case, it is a journey fit to organize the best of our abilities as we strive to stake our enterprise’s flag into the summit amongst the clouds.

This was originally published on the Juniper Networks blog and ported to my new blog here for posterity.