#MesosCon 2015

Apache Mesos has the ability to run on every private and cloud instance, anywhere. In this talk, Thomas Rampelberg (Datacenter Developer at Mesosphere) will trace Mesos from its origins as a “fine-grained resource manager” at UC Berkeley to becoming a full-fledged distributed kernel for resource management across the datacenter and cloud. He will explain the momentum behind the “single computer” abstraction that has put Mesos at the center of one of the most exciting architecture shifts in recent information technology history. Rampelberg will explain the role of Mesos in unifying other frameworks, share examples of how Mesos is empowering the new class of “datacenter developers” to program directly against datacenter resources, and draw correlations to how the Linux kernel revolutionized the server industry. 

Opentable has been using Apache Mesos for production workloads and for running critical parts of their production services for more than a year. Not only did Mesos help deploying resilient / elastic standalone applications and services , but also the distributed / fault-tolerant frameworks like Apache Spark for Data processing and machine learning. Mesos enabled Opentable to run multiple distributed applications across the same infrastructure at scale. Pablo will tell the story of how Opentable started with Mesos, the pain points of dealing with an hybrid Mesos + non-Mesos environment and how to survive in the transition.

eBay Inc. has embraced Jenkins for it’s continuous integration and delivery processes. It was challenging to efficiently use the infrastructure resources to accommodate the continuous integration process at the scale of eBay. Apache Mess and Marathon have made it possible for eBay to build a highly scalable, available and self healing processes, with efficient utilization of resources. Along the journey, eBay had challenges and overcome them effectively, and contributed back the solutions to the open source community with contributions to Jenkins framework for Mesos.

Using schedulers like Marathon and Aurora help to get your applications scheduled and executing on Mesos. In many cases it makes sense to build a framework and integrate directly. This talk will breakdown what is involved in building a framework, how to-do this with examples and why you would want to-do this. Frameworks are not only for generally available software applications (like Kafka, HDFS, Spark ,etc) but can also be used for custom internal R&D built software applications too.

Because Mesos abstracts the physical location of containers, network connectivity, service discovery, and network security can be quite challenging. In this talk, we will review the networking features necessary to make Mesos practical in large scale, multi-tenant clusters. Specifically, we will describe how to assign a separate IP address per container, how to isolate containers on demand with a distributed firewall, and how to automate service discovery by integrating Mesos with DNS. We will demonstrate the first implementation of these features that uses Project Calico, a pure layer-3 approach to virtual networking for scalable datacenter.

When it comes to hardware, most companies underutilize it; sometimes letting it lying idle for days without activity. When we started the project to build a staging environment for PayPal, we had the opportunity to design an environment, which is dynamic, on demand, and highly available. Through Mesos and the Aurora scheduler, we are able to deploy 200+ VMs simultaneously and create a mini PayPal environment for developers to connect from their VMs, desktops or mobile devices. This presentation will cover how we are reducing operational cost of hardware and building a scalable environment, which is improving the efficiency for PayPal developers. 

The promise of Mesos to create a datacenter-centric resource pool is a great idea. However, a decade old company with many different types of existing systems and unique workloads can't simply mass migrate to Mesos. That doesn't mean Mesos can't pay a huge part in unifying disparate workloads that construct existing systems while opening the door to new Mesos-centric development.

In this presentation, Steve will cover 3 very different use cases that are being unified under Apache Mesos including: Automated builds of existing code artifacts, automated builds of AWS AMIs via packer in EC2, and rolling deploys & supervision of Docker microservices across environments. The addition of Docker on Mesos has enabled this, but that doesn't mean that all Mesos slaves should be identical. See how the addition of localized support services on slaves can compliment your Docker on Mesos use cases.