968e67547a
- Revised "What is Nomad" copy - Added "Key Features" section with links to task drivers & device plugins with lift-and-shift from README - Added "Who Uses Nomad" section with users, talks, blog posts - Removed Hadoop YARN, Docker Swarm, HTCondor from comparisons - Revamped Guides section - Inserted "Installing Nomad", "Upgrading", "Integrations" as persistent in Guides navbar - Split Installing Nomad into two paths for users (one for Sandbox with "Quickstart", one for Production) - Surfaced "Upgrading" and "Integrations" section from documentation - Changed "Job Lifecycle" section into "Deploying & Managing Applications" - Reworked "Operations" into "Operating Nomad" - Reworked "Security" into "Securing Nomad" - Segmented Namespaces, Resource Quotas, Sentinel into "Governance & Policy" subsection - Reworked "Spark integration" into its own "Analytical Workloads" section
1.7 KiB
| layout | page_title | sidebar_current | description |
|---|---|---|---|
| intro | Nomad vs. Mesos with Aurora, Marathon, etc | vs-other-mesos | Comparison between Nomad and Mesos with Marathon |
Nomad vs. Mesos with Marathon
Mesos is a resource manager, which is used to pool together the resources of a datacenter and exposes an API to integrate with Frameworks that have scheduling and job management logic. Mesos depends on ZooKeeper to provide both coordination and storage.
There are many different frameworks that integrate with Mesos; popular general purpose ones include Aurora and Marathon. These frameworks allow users to submit jobs and implement scheduling logic. They depend on Mesos for resource management, and external systems like ZooKeeper to provide coordination and storage.
Nomad is architecturally much simpler. Nomad is a single binary, both for clients and servers, and requires no external services for coordination or storage. Nomad combines features of both resource managers and schedulers into a single system. This makes Nomad operationally simpler and enables more sophisticated optimizations.
Nomad is designed to be a global state, optimistically concurrent scheduler. Global state means schedulers get access to the entire state of the cluster when making decisions enabling richer constraints, job priorities, resource preemption, and faster placements. Optimistic concurrency allows Nomad to make scheduling decisions in parallel increasing throughput, reducing latency, and increasing the scale that can be supported.
Mesos does not support federation or multiple failure isolation regions. Nomad supports multi-datacenter and multi-region configurations for failure isolation and scalability.