Loom Systems was in attendance at OpenStack Summit in Boston, a four-day conference for IT business leaders, cloud operators and developers covering the open infrastructure landscape, and there was plenty to cover. We’ve known for a long time that OpenStack has arrived as the go-to public cloud approach and the OpenStack Summits around the world have proved it definitively.
Over four jam-packed days, we heard from top businesses and organizations on everything from Kubernetes, Docker, Ansible, Ceph, OCS, OpenContrail, and OPNFV. And, once again, it was clear to us how challenging it is to monitor such a decoupled & distributed environment and how valuable Loom Systems can be to businesses to mitigate problems in the entire OpenStack environment.
Here were the highlights:
Security & Visibility is Central to OpenStack
The keynote of OpenStack Summit may well have been NSA whistleblower Edward Snowden, who was interviewed via satellite link on Tuesday. Edward emphasized the need for full visibility into cloud infrastructure and complained about the opaqueness of cloud services from Amazon, Google, and Microsoft to whom businesses freely give up endless amounts of data. “You’re giving them your data and giving up control,” Edward said. Instead, Edward promoted the use of open-source technologies like OpenStack, saying they “don’t own, don’t influence, don’t control, or even shape” businesses or their data.
Of course, while OpenStack is clearly the winner when it comes to privacy, it still has its own visibility challenges. OpenStack is a complex and evolving system that generates huge amounts of log data and metrics, making monitoring a critical piece of the system. Loom Systems would be an asset to any business here running OpenStack. Since Loom Systems integrates with the entire OpenStack ecosystem, businesses would get a robust and structured monitoring tool that includes deep visibility into their applications and microservices.OpenStack plus Loom Systems would be a winning combination when it comes to both visibility and security.
CIO Cloud Strategy
One of the top talks on overall strategy came from Patrick Weeks, Senior Director of Digital Operations at GE Digital (Healthcare). CIO strategy was a top theme of OpenStack Summit and Patrick detailed GE’s transition to the cloud with OpenStack despite initially thinking they’d need to force fit their environment into AWS. In general, Patrick talked at length about the overall challenge of legacy apps, something that we know is a pain in the industry.
While it was refreshing to hear first-hand the struggles of GE when it came to OpenStack, we couldn’t help but think how much Loom Systems would’ve smoothed their path with AWS and legacy apps at large, since Loom Systems out-of-the-box integrates with both, whether or not the app is proprietary.
The top talk on Kubernetes came from eBay’s Suneet Nandwani, who spoke about how to manage Kubernetes on OpenStack at scale. It turns out eBay is one of the largest OpenStack-based clouds in the world and, as eBay has transitioned to the world of containers and microservices, Kubernetes has for them become a key platform. Suneet talked about how eBay applied their learnings from OpenStack to build a framework for managing the lifecycle of Kubernetes at scale.
Suneet didn’t go into much detail, but we got the impression that Loom Systems could’ve smoothed the way for them. Built for dynamic environments, Loom Systems is a microservice-focused monitoring solution that comfortably grows and shrinks in step with elastic services - without human intervention. When new containers spin up regularly throughout the day, Loom Systems will automatically analyze their new logs and metrics. Combined with comprehensive cross-correlation and alerts on abnormal behavior, Loom Systems could give eBay the power to anticipate and resolve everything that could impact Kubernetes performance.
Operations War Stories
Probably the most impactful operational war story came from the US Army, who brought in Major Julianna Rodriguez and Captain Christopher W. Apsey to talk about the United States Army Cyber School (USACYS) and how they transitioned from 3 borrowed servers to a 2000-core cluster backed by a 4PB Ceph array that is 100% code driven, all thanks to OpenStack.
USACYS trains and educates soldiers to defend the interests of the US and their allies at home with a combination of computer science, electrical engineering, robotics, and mathematics - and over the past 18 months, they’ve used OpenStack to execute their mission. They did a live demo of their CI pipeline with integrations into Blackboard LMS, Heat, and their asciidoc-based curriculum-as-code repository. Their OpenStack-based DevOps-oriented approach allows them to focus on teaching students while saving millions of dollars in platform expenses.
Major Rodriguez’s talk was an attendee favorite not only because it was a live demo, but also because it gave some insight into a successful application of OpenStack at an unexpected organization - the US military. For the US military, who like most businesses tap into an array of microservices, Loom Systems could be a boon. We’ve implemented all OpenStack services, including the underlying compute, storage and network infrastructures and OpenStack data-plane components, including Heat and others that the US Military relies on.
The Final Word
It was widely agreed by summit goers that the cloud is saving them money and creating a serious business impact, but we heard from many CIO and IT decision makers about the challenges of evaluating OpenStack providers and implementing a successful cloud strategy. One prospect told us that “the combinations and permutations of problems in OpenStack are endless, so traditional monitoring doesn’t work.” Architects and operators also told us that there are too many hiccups along the way to deploying OpenStack and there are too many tools purporting to solve implementation issues from disparate angles.
The overall theme was a disconnect between the promise of OpenStack and the delivery of valuable services like root cause analysis and big data analytics - the elusive unicorn of OpenStack implementation. “OpenStack is so hard because it’s still a black box,” another prospect told us, “I can’t see well into my environment to figure out the source of what’s going wrong.” OpenStack may have arrived, but too many benefits still evade CIO and IT decision makers because they don’t have enough visibility into their OpenStack environment and can’t overcome problems with implementation.
And last but not least: Many thanks to Julio Calderon, Global Senior Product Manager (Data) at KIO Networks, who agreed to share us in his experience with AI-powered OpenStack monitoring. We love you Julio!