At last week’s OpenStack Silicon Valley 2015 event, technical and financial luminaries gathered both to review OpenStack’s rapid evolution as platform and to preview the next wave of open source technology – a.k.a. containers! Kubernetes, Docker, and Magnum were the big buzzwords, with lots of hands going up during the general session when a speaker posed the question, “How many of you expect to deploy containers within the next six to twelve months?”
For organizations looking to deploy or further optimize OpenStack, here are five key takeaways from last week’s event:
1. OpenStack is poised to take a greater share of data center operations.
During the conference’s opening session, Boris Renski (Mirantis CMO and Member, OpenStack Foundation Board of Directors) cited composite analyst forecasts indicating that OpenStack will expand from 31% to 36% share of data operations between 2015 and 2018.
To achieve this growth and stave off competition from public cloud providers (a.k.a. Amazon Web Services), however, the OpenStack community will have to confront and resolve a lot of difficult challenges. Among these challenges is defining a set of standards around the containerization of applications and workloads.
However, OpenStack is more than just containers, just as it is more than a platform for virtualization. According to the OpenStack Foundation, “OpenStack includes multi-tenant security and isolation, management and monitoring, [and] storage and networking….regardless of whether containers, virtual machines, or bare metal servers are used.” The OpenStack community must harness and evolve all of these capabilities to consolidate OpenStack’s position as the de facto standard for data center operations.
2. DevOps and IT both see value in OpenStack, but for different reasons.
James Staten (formerly with Forrester Group, now Microsoft’s Chief Strategist for Cloud and Enterprise) cited key differences in how DevOps and IT perceive and measure the value of OpenStack. DevOps is attracted to OpenStack’s speed of deployment, ease of use, and time to value.
IT Operations, meanwhile, must be satisfied that broad OpenStack adoption will not compromise their functional emphasis on security, compliance, systems reliability, and cost containment. If security and compliance concerns can be overcome, OpenStack offers IT Operations the promise of increasingly cost-effective use of data center resources.
Containers potentially have a big role to play here, as the next step on the efficiency journey started by virtualization. Bridging the distinct value systems of these two constituencies – DevOps and IT -- represents the “sweet spot” for OpenStack vendors.
3. To survive and thrive in a crowded field, OpenStack vendors must offer differentiated value.
During a panel discussion on “cloudfunding,” Vivek Mehra of August Capital offered his take on the economic prospects of vendors crowding into the OpenStack field. A key point emphasized by Mehra is that customers don’t pay for open source. Rather, they pay for the differentiated features and services that vendors wrap around open source technology.
Mehra urged OpenStack users to be discerning in evaluating the relative value-add of open source vendors, and simultaneously urged vendors to clearly define to the market their value-add “above and beyond” open source definitions, tools, and resources.
4. Containers are the “latest shiny toy” that OpenStack users want to deploy onto their platforms.
Craig McLuckie (Senior Product Manager, Google) took the stage early on the first day to tee up containers as a major conference theme. Speaking credibly as an architect of Google’s container-based infrastructure and as a key sponsor of the Kubernetes project, McLuckie described how OpenStack, Kubernetes, and Docker together represent the best of what Google has architected in its own environment.
As the conference unfolded, multiple speakers described container-based environments as “GIFEE” (Google Infrastructure for Everyone Else) and dangled the promise of lightweight, portable, standardized, and easy-to-use data center infrastructure in front of receptive conference attendees. Incidentally, the “GIFEE” concept has recently coalesced into the more formal -- and formally named -- Cloud Native Computing Foundation.
According to Boris Renski, with so many vendors currently addressing the container opportunity, the current tidal wave of innovation will likely be followed by a period of standardization and inevitable shakeout. As Renski sees it, the good news for OpenStack adopters is that OpenStack is an ideal platform -- a “safe harbor” -- during periods of technological uncertainty. As a vendor-agnostic platform, OpenStack is ideally poised to support any and all “winning” container technologies and adopted standards.
5. Containers and VMs will both continue to play a role in optimizing the data center.
Throughout the conference, some heated exchanges occurred on the question of whether containers will ultimately compete -- or coexist -- with virtual machines. In multiple stage appearances, Jonathan Bryce (Executive Director, OpenStack Foundation) made a compelling case that VMs and containers will coexist for the foreseeable future.
Bryce centered his thesis on the fact that VMs and containers are not direct substitutes for one another, but rather offer distinct -- and highly compatible -- benefits to users. Most importantly, as cited recently by the OpenStack Foundation, “…containers don’t have the same security isolation capabilities as virtual machines, which means that containers cannot be viewed as a direct substitute for virtual machines.” Bryce and other speakers also highlighted the 11+ years of feature innovation and production-level “hardening” that have gone into VMs, contrasting this with the pioneering nature of current container technologies.
Many individual and panel speakers piled into this debate, with some downplaying the future role of VMs as containerization technology matures. VM detractors cited the relative ease and “lightweightness” of containers as key reasons why this technology should ultimately prevail over virtual machines. One specific speaker offered his creative vision of the future, in which users can “dial up” and “dial down” the virtualized or containerized attributes of their data centers on the fly, depending on which set of attributes best fits the specific workload or use case.
All in all, it was an insightful conference and a privilege to hear the insights of so many pioneers and leaders in the OpenStack community. The upcoming OpenStack Summit in Tokyo, Japan, promises more of the same!
To learn more about Bright OpenStack, click here.