1 / 17

Thoughts on the MS Network Research Workshop

Thoughts on the MS Network Research Workshop. Fred Baker. Terrestrial Networks for Astronomic Research. Proposed Pulsar Research Model. Internet path. Bandwidth- engineered Path: >8 GBPS. Observatory. Computation In PCs in High Schools. Servers at Swinburne.

Download Presentation

Thoughts on the MS Network Research Workshop

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. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Thoughts on the MS Network Research Workshop Fred Baker

  2. Terrestrial Networks forAstronomic Research

  3. Proposed Pulsar Research Model Internet path Bandwidth- engineered Path: >8 GBPS Observatory Computation In PCs in High Schools Servers at Swinburne

  4. One station of the proposed ~130 SKA stations in Australia 100 radio telescopes 100 sensors per telescope N2 integration of sensor feeds Built by bringing lambdas from sensors to a grid correlator Every sensor output compared to every other Results stored, original data discarded after correlation The Australian SKA Prototype

  5. Let's talk about Marketing

  6. Let’s talk a bit about marketing • Definitions: • Legacy • The old thing that works that the marketer wants to displace • Next-generation • The new thing that doesn't quite work that the marketer wants to sell • Argument style: • Emphasize interesting points (cost differences, problems with “legacy”, cool features of “next generation” approach) • Gloss over problems with new approach and strong points of the old one • Examples: • The Routing vs Ethernet Switching Wars • The Frame Relay vs IP wars • The ATM vs IP wars • The QoS Wars • The ATM vs MPLS wars

  7. The common result: • We use each technology for a purpose when it makes sense to use it • How these are seen today: • Tools in the toolbox • Not competing technologies

  8. Circuit-switch vs packet-switch question • Variation on the Routing vs Ethernet Switching Question

  9. Layer cake in the network Above Transport Transport “Network of Networks” Internet Intranet Link Layer

  10. IP Routing • Internet Layer • Used when • Connecting things that one wishes to manage the connection of • Crossing administrative boundaries • Optimizing routing • Organizing networks for maintenance • The service: • Isolation of domains of control for administrative purposes • Conscious connection of domains across the administrative boundary

  11. Ethernet switching and packet circuit switch technologies • Intranet Layer • Used when • Connecting things that one wants to treat as connected • Obscures routing • Simplifies installation • "Just works” • The service: • Circuit Switch delivers a single common service: • Point to point connectivity, potentially on demand • Administrative bounds at at a higher layer at endpoints of the circuit • Ethernet switch interconnects groups of end systems

  12. Lambda switching: • Intranet layer (a form of circuit switch) • Used when: • High capacity is required • Within an administrative domain • Breaking out a lambda is justified • Scaling of routing is not required Very reasonable place for circuit switching Circuit Switching? Not here!

  13. The greybeard speaks

  14. The arguments between Packet Switching at the Internet and Intranet layers, and Lambda Switching: • Artificial • Often essentially political • My strong suggestion: • For routing, community should use whatever technology meets its needs in each part of the network • The community should refrain from trying to force one solution to meet all needs • Make sure that your solutions meet the perceived needs not only of the users, but the operational staff that will be supporting them

  15. The place of per-flow routing and management • What ISP wants it? • Cost in telephone system largely related to micromanagement of circuits (calls) • There is a reason ISPs prefer management of aggregates • There is a reason local calls are “paid for”, and national mobile telephone networks simply sell minutes • Appropriate to large volume data flows that impose a separable cost to the network, such as perhaps lambdas

  16. Network management architecture • Network Management is something the industry has no idea how to do • We manage configurations of devices and systems • We monitor their behavior • We try to diagnose faults, with mixed success • Good suggestions that meet commercial needs are very welcome • Has to address real network requirements • Not just education or enterprise • Not just small ISP

  17. Thoughts on the MS Network Research Workshop Fred Baker

More Related