Cluster Resource Management:
A Scalable Approach

	Ning Li and Jordan Parker
	CS 736 Class Project

Outline

	Introduction
	A Scalable Approach: Hierarchy
	Results
	Conclusions
	Questions

Why Study Resource Management?

	Clusters have become increasingly popular for large parallel computing.
		Web Servers
	Clusters are becoming increasingly large to the order of thousands of nodes.
	Clusters are providing multiple services.
	Hard to evaluate
		Bad is easy to determine
		Good is much harder

Resource Management Example

	4th Node Services only B
	Poor Management
	Ideal

Clustering Goals

	Scalability
	Reliability
	High Performance
	Affordability

Related Work

	Proportional-Share
	Cluster Reserves

Related Work: Approach Differences

	Our Goal: to provide a scalable solution for resource management.
	Other work focused primarily on just having good management
		This often meant 1 manager for all the nodes
		Clearly this could present a scalable bottleneck
	Effectiveness: Other solutions probably better for smaller clusters, we hope to be better for large (>1000 nodes) clusters.

Outline

	Introduction
	A Scalable Approach: Hierarchy
	Results
	Conclusions
	Questions

Hierarchy: A Scalable Approach

	Hierarchical Management
		Nodes service jobs
		Managers facilitate resource management

Banking Algorithm

	Goal
		Determine best allocation given previous usage
	Primitives
		Tickets
		Bank accounts
			Deposit / withdraw tickets
	6 Steps

Banking Algorithm

	Step 1: For each service class on each node
		Deposit unused tickets
	Step 2: For each service class on each node
		Reallocate service class
			Full utilization: Allocation = usage + k
			Under utilization: Allocation = usage - k

Banking Algorithm Cont.

	Step 3: For each service class
		Compare total allocation to desired
			Subtract from over-allocated
			Add to needy & under-allocated
	Step 4: For each service class
		Deposit / Withdraw
			If still over-allocated withdraw
			If still under-allocated deposit

Banking Algorithm Cont.

	Step 5:
		Withdraw and allocate
			Reward the needy nodes
	Step 6:
		Done, clear the bank accounts

Reliability

	Bottom-up Manager Replacement

Outline

	Introduction
	A Scalable Approach: Hierarchy
	Results
	Conclusions
	Questions

Results

Implementation Details

	Simulations via The NS – Network Simulator
		Low bandwidth 10Mbs communication network
		UDP for lower server overhead
	Assumptions
		Node level resource management works ideally

Test 1: Overview

	4 nodes – 3 services – 60/30/10 Allocation
	4th node receives all of 3rd class’s requests
	Steady Workload

Test 1: Data

Test 2: Overview

	100 nodes – 3 services – 60/30/10 Allocation
	nodes 1-30 receive all of 3rd class’s requests
	Steady Workload

Test 2: Data

Test 3: Overview

	100 nodes – 3 services – 60/30/10 Allocation
	nodes 1-30 receive all of 3rd class’s requests
	Dynamic Workload

Test 3: Data

Test 4: Overview

	100 nodes – 3 services – 60/30/10 Allocation
	nodes 1-30 receive all of 3rd class’s requests
	Steady Workload
	Reporting 1/5
		Nodes every 0.3 second
		Managers every 1.5 seconds

Test 4: Data

Test 5: Overview

	900 nodes – 3 services – 60/30/10 Allocation
	nodes 1-300 receive all of 3rd class’s requests
	Steady Workload

Test 5: Data

Outline

	Introduction
	A Scalable Approach: Hierarchy
	Results
	Conclusions
	Questions

Conclusions

	Benefits of an hierarchy
		Scalable
		Reliable
		Geographic Applications
	Implemented a new management scheme: Banking
		Comparable Results
		Improved Scalability

Conclusions

	Clusters are sensitive to small policy changes
		Clusters are built for specific workloads
		Their performance is important and small changes have significant impact
	No scheme is universally applicable
	Future Work
		Real system implementation
			Real Workloads
			Real node level resource management
		More steady performance

Outline

	Introduction
	A Scalable Approach: Hierarchy
	Results
	Conclusions
	Questions

Questions

Related Work: Proportional-Share

	Stride Scheduling
		Ticket based and similar to lottery
	Scale
		Randomly query k nodes to find best allocation
	Different Application
		Condor-like resource allocation/applications

Related Work: Cluster Reserves

	Resource Container Schedulers
	Constrained Optimization Algorithm
	Scale
		Centralized single manager

Hierarchical Cluster Reserves – Version 1

	Modify Cluster Reserves optimization algorithm
		Use it when manager manages nodes
		AND when level_n+1 manager manages level_n managers.

Hierarchical Cluster Reserves – Version 2

	Cluster Reserves optimization algorithm
		Use it when manager manages nodes
		Don’t use it for upper level managers
	Modify the manager to manager reporting
		Lie to the algorithm