 Welcome to our talk, an introduction to magma. My name is Phil Ritter and I'm a consultant working with Facebook on the magma project. You probably heard during today's keynote that the open infrastructure foundation is collaborating with tips open core network project group and the open air interface software alliance to bring magma to an independent open source project and foster its development. As the leading open source wireless data core. Today's talk is intended to be a high level introduction to the magma project. It is not a technical presentation. We will be doing a technical deep dive session on magma at a later date. Before I start to discuss the magma project itself. We need to talk about the drive behind magma's development. Magma's mission is to bring more people online by enabling service providers with open, flexible and extensible carrier grade networks. The internet has become a crucial tool for improving livelihoods and magma strives to facilitate bringing connectivity to places where it has not generally been available. Magma exists to enable operators who are creating and extending wireless networks to simplify their deployments and their life cycle management, while providing software that operates with a high degree of assurance. The magma project strives to speed innovation and build strong support networks through a vibrant community led project. So what is magma magma is a distributed and scalable wireless data core. It delivers converged core supporting the major elements of a 4G EPC and managed Wi-Fi service. Support for 5G using the services based architecture is currently in development. The magma core is centered around an access gateway, which delivers the EPC and 5G managed Wi-Fi and 5G services. The access gateway is constructed to support a wide variety of deployment models, including bare metal deployments at the service provider edge, in private clouds, and in public clouds. The access gateway can even be deployed co-hosted with cell site equipment in an embedded configuration. The access gateway is intended to be deployed in a distributed fashion with each access gateway instance supporting a small number of radio sites. Magma also includes an orchestration and management function to configure, monitor and manage a collection of access gateways. The magma orchestrator is a central control and lifecycle management function that provides a set of REST APIs that allow integration with a wide variety of operating support systems and more global orchestration systems. It collects performance data and statistics from the access gateways and makes that data available through a web interface, as well as via APIs to upstream systems. The orchestrator is typically deployed in more central locations, often hosted in public cloud. Finally, magma provides a federation gateway that provides 3GPP compliant interfaces that allow magma to be integrated into a mobile network operators larger network. These interfaces include the S6A interface to subscriber data management and the GX and GY interfaces to online charging systems. The federation gateway also provides interfaces required to support roaming. This federation gateway is easily extensible to expose other standard and proprietary interfaces as required based on a deployment's needs. All of this is delivered as open source software and is licensed under the BSD-3 clause license, a license structure that encourages use of magma in a wide variety of deployment and support models. Magma software has been developed with open and extensible cloud native principles in mind. Magma's design is modular and modern. Each of the components of magma are reasonably self-contained and easy to extend. Magma is not burdened by the technical debt of older telco systems. Magma's design is centered around a declarative state encapsulation model with internal interfaces specifying a desired state to achieve rather than specific actions to achieve that state. This facilitates simplified scaling models, reliability, and in-service upgrade and restoration. Magma was designed from day one as a converged core. Today we support 4G and Wi-Fi and are developing 5G access, but the design is extensible to support any RAP type or even wired access methods. Magma's central controller, the magma orchestrator, offers centralized management of a magma deployment. REST APIs facilitate programmability and integration into a wide variety of operator environments. This central controller also serves as a repository for state information that can be referenced for innovative future services or to facilitate rapid restoration of service after faults in a gateway node. Finally, magma was designed from the start with continuous software delivery models in mind. I know I said this is not a technical presentation, but I'd like to look just slightly deeper at one aspect of magma's design. I noted several times that magma provides a distributed wireless core. This approach was informed by the architecture of modern data center networking and SDNs. A modern data center is constructed using a mesh fabric that provides fast, high capacity traffic distribution throughout the data center. Services are deployed on server pods with horizontal replication used for adding capacity, distributing traffic throughout the data center on the mesh network. Fault domains are limited by keeping each service instance small and adding capacity through replication. This has proven to be a highly efficient and highly effective approach for achieving massive scale reliably. While not necessarily deployed in this type of data center, magma is likewise designed to keep each access gateway small and add capacity by adding additional instances of the access gateway. Each access gateway is independently responsible for policy enforcement and connection to the data network. By doing this, we keep fault domain small and have potentially limitless scaling possibilities. By having each access gateway replicate critical state information to the centralized orchestration system. We also provide an opportunity for rapid fault recovery. Should an access gateway node fail, we can restart a new instance of that access gateway and session state information is recovered from the orchestrator will allow preservation of operating state. This facilitates rapid software upgrade opportunities. When coupled with continuous software delivery, upgrades can be test loaded on one or a small number of gateway nodes in a canary test upgrade model before being rolled out to the rest of the network. This distributed model provides significant advantages over the monolithic deployments found in traditional implementations of 3GPP systems. It avoids the traffic concentration that leads to very large network elements that are too big to fail and difficult to upgrade without impacting large numbers of radio sites and subscribers. It allows deployment of more rationally sized platforms with confidence that growth can be achieved smoothly. And it limits the exposure to the inevitable faults that occur in any computer or communication system. With contributions from Facebook connectivity and the open air interface software alliance, as well as leadership from TIP, OIF and OSA, the independent magma project starts out as a very active and dynamic software project. Over the last four months, it has averaged over 250 commits a month from more than 45 authors. More recently, contributions have started coming from authors affiliated with other organizations, including significant contributions from authors affiliated with ACL digital, FreedomFly, WaveLabs, and Rattasis. With the migration to an independent project and support from OIF, OSA and TIP, we're looking forward to growing this community and seeing significant contribution from additional authors. Magma's development has been guided by a deployment bias. This means that features and functionality that have immediate deployment opportunities are done first. Rather than focusing on a list of features that define some abstract notion of being complete, magma's mission is to bring more people online. And that has caused us to prioritize real use cases where service can be delivered to people not well served by the existing marketplace of wireless products. This has resulted in prioritizing fixed wireless access and supporting deployments with network operators in rural and developing locations around the world. Working with traditional MNOs as well as Wisps and other more innovative leaders or providers has allowed magma to focus on the basics of wireless connectivity, really, and doing them really well, concentrating on ease of deployment, reliability, upgradeability, and the distributed core. Magma has also developed a robust managed Wi-Fi service that allows network operators to better integrate Wi-Fi with their existing core. We believe that converged multi-access core is the best way to support a reliable broadband connectivity and facilitates bringing people online more quickly. Currently, we're focusing on the development of features supporting private LTE and 5G services. The magma's distributed AGW approach is ideal for the small islands of coverage and local breakout required by private LTE and private 5G deployments. Magma's small footprint provides an opportunity to embed the core functions into radio nodes, while the centralized management function and open APIs permit integration into the operation and management systems of the private network operator. Looking forward to increasing features and capabilities in magma, we see opportunities to provide network expansion to existing mobile network operator networks, and ultimately to be able to offer a rich set of services that can support a full broadband deployment with magma. Over the last few years, we've seen many open source projects come and go. What does it take to be a successful open source infrastructure project? Recent examples of projects that have come to dominate their problem domain are OpenStack and Kubernetes. Both projects launched with significant competition, OpenStack with CloudStack, Eucalyptus, and OpenNebula, and Kubernetes against Swarm, Mesos, and Nomad. What is it that ultimately made them successful? We believe the key characteristics of these successful projects are their API-centric. They're built on a set of open and extensible APIs that allow third parties to add value with both open and proprietary components. The purposeful cultivation of an active developer and user community with transparent governance. And each project had a strong, benevolent leader, one with staying power, who would support the project for the long haul. We believe magma has all of these same characteristics. Magma is a modern software project free of legacy telecom code. It's designed around open APIs both at its edge through the orchestrator and in its internal interfaces. Its internal state management approach supports ease of extension and innovation. And we have intentionally chosen the rather permissive BSD3 clause license framework to encourage adoption by both open source and proprietary users. The migration of magma to an open source project with open governance and open contribution guidelines creates an environment to grow the development community. Magma is already very active for a project of its youth and size. Over the last 90 days, seeing over 600 commits from 46 authors representing five different organizations. In an early September, the Open Air Interface Software Alliance announced that they were freezing development of their own MME platform and would be using magma software for their platform. Finally, magma has commitment from Facebook connectivity to continue their contributions to magma. The initial maintainers curating contributions to the project are Facebook connectivity and OSA engineers. And the OIF tip and OSA have together agreed to lead the independent software project. We believe this creates an ideal environment for magma to thrive and fulfill its mission of bringing more people online to a connected world. I wanted to finish up today with a call to join us to become part of the magma community. We welcome people to download magma and try it out to raise issues and offer contributions. We also welcome partners who want to engage deployments of magma or provide services to users of magma. Simple ways to join us are listed here. You can visit our website and explore our documentation and other information about magma. You can find our code on github at github.com slash magma. You can join the discussion on our Slack channel at the link shown here or find the link on the community page of magmacore.org since this link would be rather difficult to copy. Or you can participate in our first magma developers conference to be held in late January 21. Details are yet to be announced. Please watch the blog section of our website for an announcement soon. Thank you for your time today. I'm going to switch to a live Q&A session where I'll be joined by one of the key contributors and maintainers of magma, Mr. Ulos Kozot of Facebook Connectivity.