D2iQ Mesosphere vs. SUSE Rancher

Mesosphere is well suited for orchestrating workloads. It supports Docker as a container as well as support others. It is highly suitable for running resilient and auto recovering big data/application containers. Mesosphere has proven time and again to be production ready at a massive scale. It supports native single button/API call scale up and scale down and supports various deployment patterns like Blue-Green and others.

Incentivized

SUSE Rancher as a management tool becomes useful on a larger scale. Small deployments not so much. If someone also requires Kubernetes capacity or storage, Rancher is an excellent choice. Also, without Kubernetes' skills, it is unlikely that Rancher deployment is going to be a success. Then again if someone else is managing your Kubernetes capacity, setting up the software's capacity will yield greater control. Rancher is not a very integrated solution similar to others in the market.

Incentivized

• Deploying mesosphere and friends (e.g. marathon)
• Deploying applications (e.g. Cassandra, Jenkins, Spark) on to mesosphere
• Providing value add components such as velocity, and marathon-lb

Incentivized

• Public and private cloud infrastructure providers based on K8s CAPI
• REST API that can be used to integrate company services with Rancher
• GUI that is easy to learn and use in daily operations
• Builtin GitOps automation solution based on Fleet project
• It is fully open source

• Setting up is a bit of a hassle, especially ZooKeeper state management and mesos and marathon quorum.
• Occasionally, I observed some failures when deploying something onto Marathon. Logging or detailed error reporting can help.
• Stale containers and inconsistent states resultant of the cluster failure are hard to solve and need a complete system restart to get it back to normal state.

Incentivized

• No possibility to snapshot Projects. You can snapshot and restore the whole Kubernetes cluster, but not a Project or Namespace. For this, you have to use external tools.
• You cannot detach the Rancher-created Kubernetes clusters from Rancher management.

The usability and user experience are good in general, although sometimes some errors can cause confusion, especially for those users who are not experts.

Use cases can be complex hence support as well. Problems have been solvable, but not always easily. It's great that there is support!

Incentivized

I happen to like mesosphere because it integrates well with a Jenkins based workflow, Deis is a little more Heroku like and it's not clear how to fit that model into a continuous-integration process. Kubernetes has also been criticized for being complicated.

Incentivized

As we use only AWS EKS Clusters originally we were using the AWS Console and CLI but that is too limited in scope. Also, we were using AWS IAM roles to provide access to users but that was lots of extra work to have them integrated into SSO while on Rancher we have just connected our GitHub login with the Alfresco organization and that uses, in turn, Okta for SSO so provisioning for access is automatic for any developer who has been assigned to GitHub.

Incentivized

The investment for small environments is quite significant. There has to be a compelling case to enhance the areas where SUSE Rancher brings in value to make such a financial leap. There is also a free version to test the value propositions, which will help support the user's buying decisions. More clusters, more volume, more tasks and more complexity in the environment equals more value that Rancher can provide.

Incentivized

• I see mesosphere as having a positive impact overall on the industry trending Docker and containers in general.
• Seeing how mesosphere helps and simplifies things for the developer and ops, it is definitely a game changer.
• Native support of on demand scaling up and down as per the need is one of the best features.

Incentivized

• Shortens "Time-to-Market" factor for new business applications or implementing new functionalities. From 1 to 50 microservices-based business applications in 6 years.
• 24/7 availability, generates more money. There are many infrastructure components that are regularly powered-off for maintenance or upgrade, bur we rarely are turning off our downstream Kubernetes clusters where our business applications lives.
• Single Point of Contact with platform maintenance and development Team, eases implementation of new business applications