The Meltdown and Spectre vulnerabilities are proving difficult to fix, and initial experiments suggest security patches will cause significant performance penalties to HPC applications. Even as these patches are rolled out to current HPC platforms, it might be helpful to explore how future HPC systems could be better insulated from CPU or operating system security flaws that could cause massive disruptions. Surprisingly, most of the core concepts to build supercomputers that are resistant to a wide range of threats have already been invented and deployed in HPC systems over the past 20 years. Combining these technologies, concepts, and approaches not only would improve cybersecurity but also would have broader benefits for improving HPC performance, developing scientific software, adopting advanced hardware such as neuromorphic chips, and building easy-to-deploy data and analysis services.
This new form of “Fluid HPC” would do more than solve current vulnerabilities. As an enabling technology, Fluid HPC would be transformative, dramatically improving extreme-scale code development in the same way that virtual machine and container technologies made cloud computing possible and built a new industry.
Open source software (OSS) marketing today is unique: it’s a process of co-creating and co-executing a marketing plan with an entire community—developers, end users and vendors. This makes it distinctly different than most traditional technology marketing efforts, which generally focuses on business decision-makers exclusively.
OSS marketing has evolved since the emergence of OSS in the 1990s and early 2000s. It now includes the community at every step of the life cycle (a fact that shouldn’t surprise anyone familiar with open source communities’ expectations for inclusivity). It also outperforms traditional command-and-control approaches to marketing, because it’s built on the strength and reach of project communities. A community that’s bought into a marketing program will generate far more content than a marketing team alone can.
We aim here to describe a process for inclusive marketing that any technology marketer can apply to increase impact.
The evolution of OSS, from commoditizing established markets to innovating in brand new markets, means that the approach to marketing OSS today also must evolve. Two models help explain how different marketing is with today’s innovative OSS versus marketing OSS in established markets.
Key differences
OSS turned many tenets of commercial software development upside down. Open source marketing upends many assumptions of commercial software marketing in much the same way.
For example, commercial marketing teams often measure success with leads. They focus on the buyer persona. OSS marketing, in contrast, seeks to appeal to many more, and more diverse, stakeholders. Developers, maintainers, vendors, and end users are all critical to sustaining the OSS marketing effort. You can’t visualize engagement with these stakeholders in a funnel or measure it with ROMI. But without their engagement, OSS marketing efforts will fail.
Open source today
Enterprise OSS started out as the business of creating open versions of established proprietary software products. The canonical examples are operating systems (Linux, BSD), databases (MySQL, PosgreSQL) and web servers (Apache).
These successes demonstrated the advantage of inclusive development to accelerate innovation. Today, most ground-breaking software originates as open source. In business blockchain, a new category, Corda, Ethereum, Hyperledger Technologies, and Quorum are all open source. In cloud, OpenStack, Cloud Foundry, and Kubernetes are open source. Several leading AI solutions are also open source.
The evolution of OSS, from commoditizing established markets to innovating in brand new markets, means that the approach to marketing OSS today also must evolve. Two models help explain how different marketing is with today’s innovative OSS versus marketing OSS in established markets.
The first, depicted in Figure 1 below, is a generic Buyer (or in the case of OSS, Adopter) Journey marketing model. Adopters become aware of a solution, form a favorable impression, and try it. Then they adopt it—first in a limited way, then broadly. Finally, with satisfaction, they refer it to others and the cycle repeats. This kind of adopter-centric approach to marketing intersects ideal adopters (developers, vendors, end-users) on their preferred channels with useful content across their journeys.
Figure 1 (Adapted from Duct Tape Marketing)
When MySQL emerged, the entire enterprise market already used databases, so no awareness-generation activity was necessary; people already understood the uses or benefits of using databases. MySQL could focus instead on convincing users that OSS wasn’t only for geeks; indeed, many enterprises were already using MySQL without IT management’s knowledge. These users saw the same, if not superior security, performance and scalability results.
Marketers didn’t need to help potential users understand how databases could be used; they needed to persuade them that an open source database was safe to use. (As Stephen O’Grady writes in “The New Kingmakers,” free access to OSS initiated this phenomenon of developers becoming more involved in enterprise technology acquisition and consumption decisions.)
So enterprise decision-maker marketing programs for MySQL differed substantially from, say, Hyperledger. Initiated in 2016, Hyperledger is an open source effort to advance cross-industry blockchain technologies. Hyperledger marketing activities therefore focused much more on a different kind of activity: introducing people to an entirely new product category. Figure 2 offers a high-level contrast of marketing approaches needed for marketing to enterprise decision-makers in an established category versus an entirely new one.
Figure 2
To examine the second model, let’s look at the technology adoption curve. In the mature database market, for example, everyone was already using databases when OSS options emerged.
Conversely, today’s OSS innovations in new markets follow the S curve seen in Figure 3 below. Marketers must tailor programs to appeal to the appropriate segment in sequential order, beginning with innovators and early adopters.
Figure 3
As OSS has moved from commodity to new markets, the importance and complexity of marketing has increased. An open approach is now necessary for addressing that complexity.
Why consensus matters how to build it
Open source marketing builds consensus on objectives, communications, and content strategy across varied project stakeholders—rather than mandate marketing objectives or plans from the outset. Through this engagement in consensus building, the community takes on a sense of ownership for marketing. This is essential to ensure that the community contributes actively to the expanded set of activities across the adopter journey that marketers must account for when marketing innovative solutions in new categories.
Consensus relies on shared goals and values and broad agreement. Working toward consensus helps clarify, refine, and crystalize shared values and goals. It reminds all parties that they have a voice in shaping those goals and values. That’s why the path to achieving (near) consensus matters. Seeking opinions, even critiques, invites the community into the work. It’s unlikely you’ll achieve a state of complete agreement among all parties, but that’s not the point, nor should it prevent a project from trying to reach consensus.
A first step toward consensus-building is issuing a community survey to develop a shared understanding of outreach priorities. Respondents advise on key audiences, how their success is measured, where they go for trusted information, and important use cases they’re seeing.
Participants should include developers, maintainers, and project founders (who have essential insights), community managers (and others who know the project ecosystem well), and business executives. Users might reinforce a strategy or refine it. Consider involving a few industry influencers, analysts or reporters who may bring different perspectives as well.
Follow Simon Sinek’s advice and start with “Why” to draw guidance for the marketing strategy from your community. Our first survey question is “What excites you most about this project?” It’s critical to hear directly from developers about motivations, aspirations, emotions that form the backbone of the project and project community.
Responses will help project marketers align outreach with the project’s growth strategy in an open and transparent way. A next step is working closely with this core group of stakeholders to co-develop the program. Formalizing this collaborative process with an evangelism working group or messaging committee sustains engagement, ensures the project benefits from the community’s expertise, and creates shared ownership of the program.
Be willing to be wrong
Several truisms about open source now almost seem cliche, such as the superiority of “the wisdom of the crowds” and the benefits of “standing on the shoulders of giants.” More than cliches or truisms, we treat these as operational directives.
One can see the wisdom of the crowds play out in the number 80, for example: the average number of comments on a Node.js pull request tagged as “significant.” Eighty backs-and-forths, 80 times to build off of another person’s thinking. Eighty times to be a little wrong and a little right. Quality code emerges through the interaction of the community.
We apply this same principle when building community-based outreach strategies. We’ve learned that the best solution will emerge through dialogue. Taking risks and being willing to change one’s mind stimulates dialogue. The strategy is already in the community—your job is to find it, help shape it, and leverage the community to amplify it.
What success looks like
We follow this process whenever possible and have seen it bear fruit in ways all marketers can appreciate. The community that helps create the marketing strategy helps execute it.
We see this with Hyperledger and the Node.js Foundation, among other projects. Corporate members and developers co-create the strategy, contribute valuable content, and carry the project’s messages to their own communities and users.
One community-generated content example is the Node Foundation’s Enterprise Conversation podcast series. This program highlights how companies like Netflix, Lowe’s, SkyCatch, Twitter, and Google use Node. The massive Node community also engages in the Node Collection on Medium. Three volunteers edit this blog, and more than 30 volunteer writers drive more than 2,000 average unique daily visitors.
Hyperledger too sees excellent content contributions from its large community. The Developer Showcase blog series averages one community-contributed post per month. Due to the popularity of the series, Hyperledger now publishes two to three posts per month. The project’s members also delivered more than 40 talks in Hyperledger’s conference booths in 2017. Conference attendees want to know how companies use blockchain, so the community demystifies business blockchain with first-person accounts. Hyperledger also deployed technology experts from the community to speak 40 times at conferences, meetups and with the media in 2017.
Up-front engagement with the community to design the marketing strategy ensured robust participation such as this.
Community engagement strengthens ecosystems
Given these clear benefits, why not embrace community-driven marketing?
The kind of open marketing we’ve described here is far more than decision by committee. Consensus-building and harnessing the power of the full community are investments that deliver outsized returns in engagement and growth over time. They help ensure the project’s communications and outreach programs are innovating as rapidly as the technology is. And by tapping the knowledge, enthusiasm, and capabilities of the full community, open source marketing directly contributes to project sustainability.
Community-inclusive OSS marketing outperforms siloed and closed approaches. It results in high-quality, authentic, community-generated messages and content that adopters and key stakeholders value. Embracing this approach requires a change in thinking. Most important among the changes is letting the community take the lead.
Sarah Conway and Greg Wallace are co-authors of this article.
Sarah Conway – Sarah is Sr. Director of PR Services at the Linux Foundation where she works with several large, growing open source communities like Kubernetes and Node.js to create high-impact communications programs. Sarah has been involved in open source since 2004 and joined The Linux Foundation five years ago. She enjoys working with entrepreneurs and developers passionate about telling their open source success stories.
The open-source movement has gained momentum over the last few years. So much so that The Linux Foundation recently formed the LF Networking Fund (LFN) in an effort to place multiple open source networking projects under a single umbrella. These types of projects allow virtually anyone to make modifications, and potentially improve, software code through a process called upstreaming. Given the numerous open source projects available, however, choosing one to contribute to can feel overwhelming. To simplify matters, the following — though far from an exhaustive list — highlights some noteworthy open source projects.
Hyperledger
The Hyperledger Project was originally announced in 2015 by The Linux Foundation as an open source project to advance the blockchain digital technology for recording and verifying online transactions.
Linux is built for tinkering and experimentation, which means it’s always morphing and changing. New distros are popping up all the time, because all it takes is a little bit of determination, time and effort to create a custom operating system.
Not all of them hit the mark – there are stacks of Linux distros that have seen little to no action, and we’re almost certain that some have been released and never installed by anyone other than their creator.
So let’s look at our pick of the five distros moving up swiftly through the ranks in 2018.
The idea of a service mesh is still a fairly new concept for most people, so it may seem a little funny to already be talking about its history. But at this point Linkerd has been running in production by companies around the world for over 18 months — an eternity in the cloud-native ecosystem — and we can trace its conceptual lineage back to developments that happened at web-scale companies in the early 2010’s. So there’s certainly a history to explore and understand.
Before we dive in, though, let’s stay in the present a bit longer. What is a service mesh, and why is it suddenly a hot topic?
In this post, I will try to compare and contrast Julia, R, and Python via a simple maximum likelihood optimization problem which is motivated by a problem from the credit risk domain and is discussed in more detail in this post.
TL;DR
For such a simple optimization problem, R, Julia, and Python/SciPy will all do a competent job, so there is no clear winner. However as noted by Julia discourse member ChrisRackauckas
if you wish to solve only a single optimization problem and that problem takes <10 seconds, then Julia’s long initial compilation is something you want to avoid. For long enough problems, or for solving multiple optimization problems, the compilation time is not noticeable.
Kelsey Hightower, Developer Advocate at Google, kicked off the KubeCon + CloudNativeCon event in Austin with an opening keynote in which he demonstrated Kubernetes’ ease of use with the help of his smartphone. Apart from commending the audience for making Kubernetes the boring-in-a-good-way framework it is today, Hightower also warned about how Kubernetes should not be considered the end game, but a means to an end.
In his talk, Hightower first addressed the misconception that Kubernetes is difficult to install. He did so by installing an eight-node Kubernetes cluster in less than two minutes just by giving verbal instructions to the Kubernetes Engine assistant through his smartphone, thus proving tha Kubernetes generally gets out of the way quickly.
Second, Hightower explained that using kubectl for everything on your cluster is mostly not necessary. Developers should not be interested in having kubectl installed on their laptops, for example, as the workflow developers are most interested in, that is, the steps that take some code from modification to production, passing through testing and QA, in no moment require access to kubectl.
Hightower drove this point home with another demonstration in which he cloned a Git repository containing a dockerized app, created a new branch, modified the code and then pushed it to testing in a staging environment.
He explained that if you pushed to any branch except master, you were really asking to deploy the code in staging. If you tagged your code, you wanted it to go into QA. Once you had gone through those steps, you could deploy it to production by pushing to master. He went through each of these steps without once having to resort to Kubernetes-specific tools like kubectl.
Hightower also showed that, by providing monitoring tools such as a graphical dashboard and logging tools, the need for command line Kubernetes tools was further diminished.
Hightower then capped his demonstration by once again using his smartphone to talk to the Kubernetes Engine assistant and deploying his dockerized application on 10 replicas.
ELC, now in its 13th year, is the top vendor-neutral technical conference for companies and developers using Linux in embedded products. Attend to collaborate with the technologists working on the latest in embedded systems, frameworks and applications.
The co-located OpenIoT Summit is a technical conference for developers and architects working on industrial IoT. Attend to increase your technical knowledge needed to deliver smart connected products that take advantage of the rapid evolution of IoT. This is the only IoT event focused on the development of open IoT solutions.
Attendees have access to all sessions and activities at both events.
Sign up for ELC/OpenIoT Summit updates to get the latest information:
In addition to great speakers and sessions, attendees have access to the following activities and events:
Birds of a Feather Sessions: Unconference sessions let you work openly and collaborate with other top leading professionals in the embedded Linux and OpenIoT fields to solve today’s challenges and create tomorrow’s innovations.
Onsite Attendee Reception:Enjoy drinks and light bites as you connect and engage with other attendees while checking out the latest technologies from event sponsors and the community in theELC Technical Showcase.
*New* Embedded Apprentice Linux Engineer Track:This new track is aimed at embedded engineers who are transitioning to Linux. The track is a set of nine seminars over three days – take as many or as few as you need. Seminars start with a 45-60 minute lecture or presentation, followed by an hour of lab time to practice relevant skills using a PocketBeagle board. $75 Registration Fee – to cover the cost of the HW kit.
Yocto Project Developer Day North America 2018:A one-day, hands-on training event that puts you in direct contact with Yocto Project developers. Developers will walk you through how to create custom-built Linux distribution for embedded devices by using layers and recipes designed to resolve incompatibilites between different configurations. Additional registration fee of $209, registration will increase on 2/28/18 to $249.
The Closing Game: A perennial favorite – the Closing game is part trivia, part pop culture, all fun, helping to close out this great event. You don’t want to miss this year’s version!
Final chance to save $150
Standard pricing closes in a few days. Register by February 17 to save $150.
LinuxQuestions, one of the largest internet Linux groups with 550,000 members, has just posted the results from its latest survey of desktop Linux users. With approximately 10,000 voters in the survey, the desktop Linux distribution pick was: Ubuntu.
While Ubuntu has long a been popular Linux distro, it hasn’t been flying as high as it once was. Now it seems to be gathering more fans again. For years, people never warmed up to Ubuntu’s default Unity desktop. Then, in April 2017, Ubuntu returned to GNOME for its default desktop. It appears this move has brought back some old friends and added some new ones.
There has never been a better time to be a DevOps engineer. Compared to traditional web stacks, containerization has dramatically streamlined the task of deploying web services such as databases, key/value stores and servers. Furthermore, container orchestration tools, like Google’s Kubernetes and Docker Swarm, enable organizations to automate the deployment and management of these containerized applications. But the tools that make life easier and more efficient for engineers can also be a gift to an attacker.
Regardless of the initial exploitation vector, an attacker’s first objective is often to gain host-level access to a target system. With that access, an attacker can leverage the system for a variety of malicious purposes — to exfiltrate data, to maintain a point of presence, to move to higher-value assets in a network, etc.
