Inferring the Topology and Traffic Load of Parallel Programs in a VM

environment

Ashish GuptaPeter Dinda

Department of Computer ScienceNorthwestern University

Overview• Motivation behind parallel programs in a

VM environment• Goal: To infer the communication

behavior• Offline implementation• Evaluating with parallel benchmarks• Online Monitoring in a VM environment• Conclusions

Virtuoso: A VM based abstraction for a Grid environment

Motivation

• A distributed computing environment based on Virtual Machines– Raw machines connected

to user’s network– Our Focus: Middleware support

to hide the Grid complexity

Motivation

• A distributed computing environment based on Virtual Machines– Raw machines connected

to user’s network– Our Focus: Middleware support

to hide the Grid complexity

• Our goal here: Efficient execution of Parallel applications in such an environment

ParallelApplication Behavior

Intelligent Placement and virtual networking

of parallel applications

VM Encapsulation Virtual Networks With VNET

VNET

• Abstraction: A set of VMs on same Layer 2 network

• Virtual Ethernet LAN

Goal of this project

Low Level Traffic Monitoring

?

An online topology inference framework for a VM environment

Application Topology

Approach

Design an offline framework

Evaluate with parallel benchmarks

If successful, design an online framework for VMs

An offline topology inference framework

Goal: A test-bed for traffic monitoring

and evaluating topology inference methods

The offline method

Synced Parallel Traffic Monitoring

Traffic Filtering and Matrix Generation

Matrix Analysis and Topology Characterization

The offline method

Synced Parallel Traffic Monitoring

Traffic Filtering and Matrix Generation

Matrix Analysis and Topology Characterization

The offline method

Synced Parallel Traffic Monitoring

Traffic Filtering and Matrix Generation

Matrix Analysis and Topology Characterization

The offline method

Synced Parallel Traffic Monitoring

Traffic Filtering and Matrix Generation

Matrix Analysis and Topology Characterization

The offline method

Synced Parallel Traffic Monitoring

Traffic Filtering and Matrix Generation

Matrix Analysis and Topology Characterization

PVMPOV Inference

Synced Parallel Traffic Monitoring

Traffic Filtering and Matrix Generation

Matrix Analysis and Topology Characterization

Infer.pl

Parallel Benchmarks Evaluation

Goal:To test the practicality of low level

traffic based inference

Parallel Benchmarks used

• Synthetic benchmarks: Patterns– N-dimensional mesh-neighbor– N-dimensional toroid-neighbor– N-dimensional hypercubes– Tree reduction – All-to-All

• Scheduling mechanism to generate deadlock free and efficient schemes

1 2 3

Application benchmarks

• NAS PVM benchmarks– Popular benchmarks for parallel computing– 5 benchmarks

• PVM-POV : Distributed Ray Tracing• Many others possible…

• The inference not PVM specific– Applicable to all communication .– e.g. MPI, even non-parallel apps

Patterns application

2-D Mesh 3-D Toroid 3-D Hypercube

Reduction Tree All-to-All

PVM NAS benchmarks

Parallel Integer Sort

Traffic Matrix for PVM IS benchmark

Traffic Matrix for PVM IS benchmark

Placement of host1 is crucial on the network

An Online Topology Inference Framework: VTTIF

Goal:To automatically detect, monitor and

report the global traffic matrix for a set of VMs running on a overlay network

Overall Design

• VNET– Abstraction: A set of VMs on same Layer 2

network– Virtual Ethernet LAN

A VNET virtual layer

VNET Layer

Physical Layer

A Virtual LAN over wide area

Overall Design

• VNET– Abstraction: A set of VMs on same Layer 2

network• Extend VNET to include the required features

– Monitoring at Ethernet packet level• The Challenge here

– Lacks manual control– Detecting interesting parallel program

communication ?

Detecting interesting phenomenon

Reactive Mechanisms Proactive Mechanisms

•Certain address properties

•Based on Traffic rate

•Etc.

Provide support for queries by external agent

Rate based monitoringNon-uniform discrete event sampling

What is the Traffic Matrix for the last n seconds ?

Like a Burglar Alarm Video Surveillance

Traffic Analyzer

Rate based Change detection

Traffic MatrixQuery Agent

VM Network Scheduling Agent

VNET daemon

VM

VNET overlay network

To other VNET daemons

Physical Host

Traffic Matrix Aggregation

• Each VNET daemon keeps track of local traffic matrix– Need to aggregate this information for a global view– When the rate falls, the local daemons push the traffic

matrix (When do you push the traffic matrix ?)– Operation is associative: reduction trees for scalability

The proxy daemon

Evaluation

• Used 4 Virtual Machines over VNET • NAS IS benchmark

Conclusions

Possible to infer the topology with

low level traffic monitoring

A Traffic Inference Framework for Virtual Machines

Ready to move on to future steps:Adaptation for Performance

Current Work

• Capabilities for dynamic adaptation into VNET

• Spatial Inference Network Adaptation for Improved Performance

• Prelim Results: Improved performance upto 40% in execution time

• Looking into benefits of Dynamic Adaptation

For more information

• http://virtuoso.cs.northwestern.edu• VNET is available for download

• PLAB web site:

plab.cs.northwestern.edu