Interactive control over a programmable computer network using a multi touch surface
This presentation is the property of its rightful owner.
Sponsored Links
1 / 13

Interactive Control over a Programmable Computer Network using a Multi-touch Surface PowerPoint PPT Presentation


  • 46 Views
  • Uploaded on
  • Presentation posted in: General

Interactive Control over a Programmable Computer Network using a Multi-touch Surface. Rudolf Strijkers 1,2 , Laurence Muller 1 , Mihai Cristea 1 , Robert Belleman 1 , Cees de Laat 1 , Peter Sloot 1 , Robert Meijer 1,2. 1 University of Amsterdam 2 TNO Information and Communication Technology.

Download Presentation

Interactive Control over a Programmable Computer Network using a Multi-touch Surface

An Image/Link below is provided (as is) to download presentation

Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author.While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server.


- - - - - - - - - - - - - - - - - - - - - - - - - - E N D - - - - - - - - - - - - - - - - - - - - - - - - - -

Presentation Transcript


Interactive control over a programmable computer network using a multi touch surface

Interactive Control over a Programmable Computer Network using a Multi-touch Surface

Rudolf Strijkers1,2, Laurence Muller1, Mihai Cristea1, Robert Belleman1, Cees de Laat1, Peter Sloot1, Robert Meijer1,2

1University of Amsterdam2TNO Information and Communication Technology


Introduction

Introduction

What are programmable networks?

When are programmable networks useful?

Interactive control over a programmable network

Demonstration at Super Computing 2008

Conclusions


Programmable networks

Programmable Networks

Applications

Applications

TCP/UDP/…

Internet Protocol

Software Objects

Virtualization

Network


Programmable networks1

Programmable Networks

Applications

Applications

Sockets

Proxies

Routing

Middleware


When are programmable networks useful

When are Programmable Networks useful?

  • When precise control of network resources matter

    • Real-time experiments using Grids (e-VLBI)(on-demand network resources)

    • Distributed applications where network performance outweighs CPU performance in scalability issues

  • Advanced, application-specific network demands, which can change over time

  • Networks that need to adapt to environmental changes

    • Sensor Networks

    • Resilient and self-healing networks


Basic issue in programmable networks

Basic Issue in Programmable Networks

  • Other than in IP networks, there is not a single end-to-end service, but rather a collection of software objects

  • How to program and coordinate such a network?

    • Configuration and setup files

    • Domain specific programming languages

  • Here, we explore a dedicated human-network interface

    • Gestures automatically translate to manipulations in individual network elements

    • Real-time, direct interaction in the form of visualization of network state

    • Enable multiple persons to manage or monitor the network simultaneously


User programmable virtualized networks

User Programmable Virtualized Networks

Packets are tagged with identifiers (e.g. string, URL) to bind application behavior to a network service

Application

Application

NC

NC

NC

AC

AC

AC

AC

AC

Architectural framework to model network behavior as part of the application domain

Applications use proxies (Network Components) to program a network service

Network cannot know the behavior in advance

Allow uploading of Application Components


Interactive control over programmable networks

Interactive Control over Programmable Networks

We have developed an experimental platform that combines fine-grained traffic control, interactive visualization and decision making into one system

An elementary control loop

Network state is visualized by monitoring ACs

UPVN takes care of translatinggestures into manipulationsof individual network elements

The human is the optimizer


Interactive control over programmable networks1

Interactive Control over Programmable Networks

20 machines (of which 12 VM and 4 end-systems and 2 controllers) connected by a 1Gbit network and separate control network


Implementation

Implementation

Other visualization, analysis or manipulation tools

Controller

Multi-Touch Table

Multi-touch Adapter Service

Visualization

Other services

Control

3

3

Monitoring Service

Compiler Service

  • Controller is merely a known peer that other peers connect to

  • Contains functions that span over multiple peers

    • 2phase commit

    • Connection monitoring

3

Low level functions (ARP discovery, traffic processing)

Adapters to middleware services or low-level control loops

Asynchronous p2p messaging infrastructure

NE

Service

2

NE Functions

1


Demonstration @ super computing 2008 austin texas

Demonstration @ Super Computing 2008 Austin, Texas


Future work

Future Work

  • Include Optical/Hybrid networks

    • Describe network resources into a semantic model (for example with Network Description Language)

    • Automated reasoning on semantic model to optimize network and distributed application

  • Interactive visualization of large-scale grids and networks

    • How to visualize and interact with millions of nodes?

    • Zooming user interface

  • Integrate network programmability in grid middleware frameworks

    • First efforts with WS-VLAM (workflow manager)

    • MPI support partly finished


Conclusions

Interactive networks combines traffic processing, visualization and control of a programmable network into one system

With interactive visualization programmable network behavior can become easier to understand

Visualization can be focused on specific problems, such as detecting bottlenecks, available resources or failures

Using a large multi-touch interface enables collaborative problem solving

Multiple persons standing around the table can simultaneously interact with the programmable network

Conclusions


  • Login