LSF란 (IBM Spectrum LSF)

01 LSF란

1. 정의

• 'Load Sharing Facility'의 약자
• IBM에서 만든 소프트웨어 제품으로 간략하게 말해 job scheduling 프로그램 (workload management platform)
• 빠르고 믿을만한 workload performace를 실행하고 비용 또한 절감되는 shared, scalable, and fault-tolerant 인프라를 생성하기 위해 다양한 IT 자원(resource)에 작업을 분배 (distribute jobs)함
• 부하(load)를 균형있게 분산하고 자원을 할당하며 해당 자원에 접근할 수 있는 기능을 제공
• provides a resource management framework that takes your job requirements, finds the best resources to run the job, and monitors its progress
• 작업 요구 사항을 고려하여 작업을 실행하기에 최적의 자원을 찾고 작업의 진행 상황을 모니터링하는 자원 관리 프레임워크를 제공
  • (Job) 작업은 항상 호스트 부하(host load)와 사이트 정책에 따라 실행됨

02 LSF Cluster란

1. Cluster의 정의

• A group of computers (hosts) running LSF that work together as a single unit, combining computing power, workload, and resources. A cluster provides a single-system image for a network of computing resources.
• Hosts can be grouped into a cluster in a number of ways. A cluster can contain:
  • All the hosts in a single administrative group
  • All the hosts on a subnetwork

2. Hosts

• Hosts in your cluster perform different functions

1) Management host

• LSF server host that acts as the overall coordinator for the cluster, doing all job scheduling and dispatch

2) Server host

• a host that submits and runs jobs

3) Client host

• a host that only submits jobs and tasks

4) Execution host

• a host that runs jobs and tasks

5) Submission host

• a host from which jobs and tasks are submitted

03 Job이란

• A unit of work that is running in the LSF system.
• A job is a command that is submitted to LSF for execution. 
• LSF schedules, controls, and tracks the job according to configured policies.
• Jobs can be complex problems, simulation scenarios, extensive calculations, or anything that needs compute power.

04 Job slot이란

• A job slot is a bucket into which a single unit of work is assigned in the LSF system.
• Hosts can be configured with multiple job slots and you can dispatch jobs from queues until all the job slots are filled.
• You can correlate job slots with the total number of CPUs in the cluster.

05 Queue란

• A cluster-wide container for jobs. 
• All jobs wait in queues until they are scheduled and dispatched to hosts. 
• Queues do not correspond to individual hosts; each queue can use all server hosts in the cluster, or a configured subset of the server hosts. 
• When you submit a job to a queue, you do not need to specify an execution host. 
• LSF dispatches the job to the best available execution host in the cluster to run that job. 
• Queues implement different job scheduling and control policies.

06 자원(Resources)이란

• Resources are the objects in your cluster that are available to run work. 
• 예시) include but are not limited to hosts, CPU slots, and licenses

참조

1. https://www.ibm.com/docs/en/spectrum-lsf/10.1.0?topic=overview-lsf-introduction
2.  
3.  
4.