What is it?
The Odin framework provides a scalable distributed environment for resilient process handling across a diverse set of agent processes and hardware types. It is based on several well known distributed open source products for both synchronization and data handling tasks. It is the framework on which several of our most popular products run to deliver seamless services and job management across hybrid and symmetrical platforms (both on-premise and in a third-party cloud solution).
What makes it different?
While many products can marshal resources from hybrid environments, a differentiating factor between those systems and Odin is scalability and resilience. Odin management clusters can be deployed with minimal hardware requirements versus other management solutions and then scaled as needed to meet client needs. Odin clusters also can have additional director elements (referred to a shards) added without interrupting the overall managed calculation fabric in real-time. Here is a quick list of the enhanced feature of Odin:
- Odin does not require elevated permissions globally in order to execute tasks
- It runs tasks through ssh (delivering implied encryption of request processing between the management shards and the devices receiving the requests)
- Odin can deliver value without installing any agent software on the devices to be managed
- Odin is based on FreeBSD and does not exhibit the traditional bloat or security weaknesses found in other operating systems
- Within the OS on which the shard is hosted Odin is operated from within a "Jail" meaning it cannot access resources it is not specifically authorized to access
- Odin is highly scalable
- New shards can be added to increase resilience and/or throughput without disrupting the job flow currently running on the fabric
- A shard can be brought down and the underlying infrastructure will repair itself until it is brought up or replaced
- A replacement shard can automatically integrate itself into the management layer with no disruption of service
- Odin can offer performance gains over traditional solutions because of its shard architecture
- Agent processes can access shards closer to them and overall load can be more evenly distributed across the framework (as opposed to a single manager which can become a bottleneck)
- Additional shards can also be provided for burst processing during high load periods (EOD processing and Quarterly/Yearly reporting)
- Components are coded in C producing the highest possible performance
- An Odin shard can be run either "on the metal", in the cloud or a hybrid of both as client needs dictate
- Odin does not require special or proprietary hardware and can be run off of commodity components making it an economical solution and saving the heavier machines for actual processing of client workloads
- Odin is self repairing so manual interventions are minimal
- The addition or replacement of an Odin shard is non-impactful in most cases
- Agents attached to the Odin cluster can reconnect to another available shard if their current one becomes unavailable (minimizing manual intervention in DR scenarios and providing uninterrupted job flow)