I'm working on an infrastructure project where our organization has several physical servers across different departments, each running its own set of applications and websites. I'm considering using Apache Mesos to optimize resource allocation by clustering these servers together. However, I have some constraints and questions regarding the feasibility of this setup:
Background
- Each department's server has unique, non-containerized applications/websites.
- The servers are under varying loads, with some being heavily used and others less so.
- The goal is to cluster these servers to maximize resource utilization without major changes to our current setup.
Constraints:
- The applications/websites are not containerized.
- Significant changes to the application source code or their operational environment are not desirable due to departmental protocols and technical complexity.
Questions:
- Can Apache Mesos be effectively implemented in a multi-server, multi-department environment to function as a unified cluster?
- Is it possible to integrate with Mesos with minimal or no changes to the source code of existing applications and websites? Specifically, are there strategies to use Mesos for resource optimization without needing to containerize these applications or alter their hosting setup?
I'm seeking insights on whether Mesos can be a suitable solution under these constraints and how it can be implemented in such a scenario.
Any advice or shared experiences with similar setups would be greatly appreciated.
Thank you!
I researched Apache Mesos, expecting it could integrate with our existing server setup to optimize resource usage without major changes to our applications. My focus was on understanding whether Mesos could manage resources at the infrastructure level, allowing our current non-containerized applications to remain as they are while enhancing efficiency. However, I'm unsure about how this aligns with real-world implementation in our specific scenario, prompting me to seek further insights.