If there's a common thread that runs through all DC/OS deployments, and these five specifically, it's efficiency. Among other things, DC/OS makes it dead simple to:
- Start, scale and productionize Docker environments
- Start, scale and productionize big data environments (notably running Spark, Kafka and Cassandra)
- Slash server and cloud footprints, resulting in lower opex
- Standardize the developer experience across cloud and datacenter environments
- Deploy new code, or entirely new applications
- Manage all of this with just a few (or less) operations staff
We're talking about deploying and managing complex systems at large scale, and doing so in a fraction of the time with a fraction of the resources. So rather than just buying resources from cloud providers, for example, DC/OS users are actually operating like cloud providers and taking advantage of the fast pace of innovation that enables.
Autodesk and Time Warner Cable
Mesosphere customers Autodesk and Time Warner Cable, as well as Cisco, presented at MesosCon alongside Mesosphere CEO Florian Leibert to explain the results they've seen with DC/OS:
Autodesk is a remarkable story for a number of reasons, including how it cut its Amazon Web Services footprint by 66 percent, is deploying new builds in under a minute, and is targeting running DC/OS across cloud providers to mitigate lock-in and give developers a standard user experience.
Time Warner Cable is running DC/OS on top of OpenStack in order to give its developers an easy way to build microservices applications on Docker containers, and to easily take advantage of big data technologies such as Apache Kafka and Apache Cassandra. Scheduled for production deployment this month, the company's DC/OS environment will span datacenters on the east and west coasts of the United States.
Another Mesosphere customer, Esri, demonstrated its forthcoming ArcGIS service that utilizes DC/OS by way of the Microsoft Azure Container Service. ArcGIS takes advantage of data systems such as Spark Streaming, Kafka and Elasticsearch, as well as the Azure IoT Service, in order to analyze and visualize geospatial data in real time. The demo culminates with an interesting use case, showing off how a geofence could trigger an interaction between connected taxi and connected rider to make sure they don't miss each other, even in a crowded location.
You can get a little more info on what led Esri to DC/OS, and how it helps them scale ArGIS to unprecedented levels, in this interview with Esri's Adam Mollenkopf:
Pioneering DC/OS customer Verizon showed off how it's able to analyze streaming data in order to address the Internet of Things. The demo—which is based on an actual application Verizon is currently developing—simulates 110 million connected devices generating 400,000 events per second and hitting datacenters on both U.S. coasts. Like Esri, Verizon uses Spark Streaming and Kafka to handle the data streams, as well as Cassandra on the backend.
And just two months into launching the project our open source DC/OS project, the platform already has at least one very cool production use case on record: Healthcare startup Wellframe is using DC/OS to simplify operations for its data-driven personal-care platform consisting of Hadoop, Spark, Cassandra and Kafka. Interestingly, Wellframe had already seen significant operational improvements from Apache Mesos, but was able to step up its pace of innovation by transitioning to the Mesos-based DC/OS platform.
All in all, the company has seen the following results:
- 1.5 months to move full stack to Mesos
- 1 month to migrate to DC/OS
- 67 percent reduction in server instances
- 33 percent cost savings on servers
- 50 percent reduction in operational overhead (engineering time)
You can read the full Wellframe blog post here.
[caption id="attachment_5564" align="aligncenter" width="800"] Source: Wellframe[/caption]
Keep DC/OS growing
Speaking of open source DC/OS, the really nice thing about these user stories is that these aren't just users of the platform, they're contributing to the DC/OS community. Whether they're Mesosphere customers, open source users and/or open source contributors, everyone's involvement matters. A stronger ecosystem means better DC/OS technology, more services in the Universe marketplace (which installs distributed systems with a single click or command) and more integrations with other popular technologies.
For example, 61 days into its existence: thousands of users have launched open source DC/OS clusters; six new services have been added to Universe—Avi Vantage, Concord, Jenkins, Kafka Manager, Spark Notebook and VAMP; several new partners have come on board, including CloudBees and DataStax; and we have added one new open source component, the Shakedown tool for testing DC/OS services.
You can learn more about why companies are excited to integrate with and develop services for DC/OS by watching this interview (shot at MesosCon 2016) with early partner NGINX. NGINX is building a secure (encrypted) and high-performance microservices architecture on top of DC/OS, which the company says is enabled by the distributed nature of DC/OS and the types of container operations it supports.
If you want to learn more about the DC/OS open source project, browse its website at dcos.io and check out the newly launched blog. You can also learn more about where it's headed and how to get involved, by watching this MesosCon 2016 presentation from DC/OS General Manager Aaron Williams.
To keep up with everything that's happening at Mesosphere, follow this blog and check out our events page to see where our engineers and executives will be speaking.