The New Stack Makers

Alex Williams discusses cloud, containers and microservices with Jason McGee, IBM Fellow, VP and CTO, Cloud Foundation Services at IBM. The discussion begins, as good discussions tend to, with a brief history. “As you’re probably aware,” Jason says, “IBM has a pretty long history in what you would think of as ‘virtualization’ and workload management technology — ‘How do we run multiple applications in shared infrastructure?’ — going all the way back to some of the seminal work around the mainframe, and even work that we’ve done on Linux.” “For a long time,” he says, “we’ve been thinking about how to package workloads, how to run workloads side-by-side on a computer system, and control isolation and sharing of resources.” From IBM’s perspective, according to Jason, these problems led to the beginning of containers. The early work IBM did 11 or 12 years ago involved looking at how to build into the Linux kernel some of their ideas from mainframe systems — AIX and other UNIX systems — and put in the primitives that allowed resource isolation between different applications.

Described as “a hackable text editor for the 21st Century,” Atom can be fully customized to do anything but also used productively, “without ever touching a config file.” A built-in package manager enables you to search for and install new packages or start creating your own — all from within Atom. Atom lets you easily browse and open a single file, a whole project, or multiple projects in one window, and can be used on OS X, Windows, or Linux. “GitHub is known for being the place for social coding, and where all of the code is hosted, but there’s so much that happens in building software before it gets on GitHub. We wanted to build a tool that would enable us to create the future of social coding,” says Brandon. “What if we could actually work together while we were building the code?”

At Capital One, says platform engineering fellow Dr. Tapabrata Pal (who goes by “Topo”), they have been doing DevOps for four years. During this time of using multiple tools, each with its own dashboard, they identified a need for a single dashboard that could view the whole pipeline in one single dashboard. Not finding anything in the commercial or open source markets, they decided to build themselves one. “It gets all the information from each individual tool, collects it in a central place and have a nice UI to show you the actual information coming out of these tools. The whole Idea is to shorten and amplify the feedback loop,” says Topo, citing the second of Gene Kim’s "Three Ways of DevOps." Topo suggests development teams who are working concurrently on isolated projects as a prime use case, in which the dashboard would signal to everyone if a build should fail, shortening the time it takes to determine the cause.

As long-time FreeBSD community member Michael Dexter suggests in his recent post, “Docker Done Right,” “FreeBSD just may prove to be the ultimate Docker platform thanks to its 15+ years of containment experience and the unrivaled OpenZFS file system.” Docker is available on FreeBSD as of June, 2015. Micahel sees Docker as “a sysadmin’s approach to software packaging.” “The sysadmin has their hands tied,” says Michael, “and can’t modify things above them and below them. So what do you do? You create a shim environment where you make it easy to port applications.” “As a platform,” he explains, “once the Linux emulation hit 64 bit, something like Docker was a matter of weeks of porting, in spare time, rather than a massive engineering feat.” Michael is passionate about FreeBSD, which, he says, “goes back directly to the CSRG and the BSD project.” “When you take a platform like that that’s ready for such things (like Docker), it just dropped into place.” “We’ve had a very clear ports tree, and application management from that perspective, so we’ve been containing applications from ports in jails for a very long time,” he says. He believes users who are lured to the FreeBSD platform and who discover how well it performs with Docker will not want to go back. “Once you spend some time with ZFS you will never want to see another hardware rate card in your life again. Once you do a ZFS send on important data, you’ll never want to use rsync again.”

You hear HP executives talk more about developers, the DevOps movement and how the company is viewing analytics and the way it works in operations. I sat down with Roy Ritthaler, vice president of product marketing at HP Software who talked about how customers are now accepting automation. As this automation movement accelerates, HP is developing a DevOps product set as illustrated in the launch of LeanFT, which we wrote about yesterday.

A recap of day one at OpenStack, live from a restaurant with a name that we are not really sure if it really is pronounced Pho Qieu. We discuss: Tesora, the infrastrucure for DBaaS on OpenStack and their reach for the top 5% of the IT market. The new networking, low-level challenges and the problems with network automation. The problem with OpenStack's demo: "The dude uploads to the cloud, the dude downloads from the cloud." DefCore: What is it? Are you listening Rob Hirschfeld? Containers and OpenStack. Apcera and AppC, the container standard. Vmware, containers and security. Open source communities: Will there be a Docker Foundation? Defense mechanisms. Is OpenStack the happy medium? Citrix joins OpenStack. What's up with that? And a last good bye from....Pho Qieu.

Erik is a 2011 Code for America Fellow and was co-founder of 140 Proof. Previously, he was system architect at Fluther and is now developer evangelist at SoundCloud. He is also an ardent first-wave Rubyist.

We have been seeing more examples of how the Go programming language’s concurrency and its simplicity are advantageous over other languages. Go’s concurrency is supported by its use of "goroutines" and channels, and in this talk from Gophercon India, William Kennedy — managing partner at Miami-based web, mobile and systems development company Ardan Studios and author of GoingGo.net and Go In Action — talks about how Go’s channels give it a consistency and predictability.

Codeclimate by The New Stack Makers

Christophe Philemotte from Brussels Belgium is the founder of PullReview. With a PhD in AI, Christophe is also Ruby and C++ developer.He founded PullReview as an automated code review for Ruby and Rails developers. In his presentation at Rubycon, India, Christophe discussed the elements of diving into the unknown depths of a project. He talked about chartering a plan to dive into a code base and read and write code then defining a goal to mapping and using the right equipments available to develop.

Bryan Helmkamp is the founder and CEO of Code Climate. He is an active member in the Ruby community and developed Code Climate to take advantage of static analysis. In 2009, Bryan received the Ruby Hero Award for his efforts. At the Rubycon held in Goa, India, Bryan talked about "4 ways to deploy a production website with rigour". He highlighted the techniques available namely Rollout, Branch deploys, Theory and Scientist that Code Climate has adapted. Bryan talked about the importance and significance of these techniques and how you can deploy intensive applications to production even without a staging process.

Sam Ramji, Cloud Foundry: The Flipside of Portability by The New Stack Makers

Here's part two of our interview with Sam Ramji who discusses what it was like to do briefings for Bill Gates about open source during a time when Microsoft was so Windows centric.

This is part one of a three part interview series with Sam Ramji, well-known technologist and the CEO of the Cloud Foundry Foundation.

A discussion on ideas which Go's designers decided to leave out of the language and the effect of those decisions on a language designed for simplicity and scale.