Internet Society recommends development of Solar-System-scale routing framework
- Reference: 1663658946
- News link: https://www.theregister.co.uk/2022/09/20/ipnsig_solar_system_routing_framework/
- Source link:
The SIG’s call was revealed in its [1]September newsletter , which detailed a late meeting of the body’s Architecture and Governance Working Group that considered current approaches to routing data traffic in space and sought to produce a recommendation toward realizing a common interplanetary network architecture.
That recommendation has now been issued and outlines four principles the Group believes are necessary for a routing framework that can serve the solar system, namely:
Autonomy
The overall goal is to populate forwarding tables or routing tables (where appropriate) without human intervention and depart from the Earth-centric management scheme
Commonality
To enable Autonomy we need a common and standardized way of the following:
The population of a Forwarding and (where routes are computed) Routing Table
Function enabling the bundle protocol agent to access a forwarding table API that answers the question: “where should I send this bundle and with what CLA/address?”, similar to ARP lookup Interoperability
Define a common to every Node set of functions
Interoperability
Inter-regional routing should be based on a common standard
Intra-regional routing could be heterogeneous, adopting methods that suit the needs of a specific environment or a mission
Inter- and Intra-region routing should be interoperable
Scalability
Architecture needs to be structured in some way (Hierarchy)
Node IDs, Addressing Schemes, and segmenting concepts need to be further explored to accommodate the scale of the SSI
The IPNSIG has already developed a technology called “Delay-Tolerant Networking” (DTN) that addresses the issue of latency. DTN tries to address the fact that in space, long distances or network nodes disappearing behind a planet create latency and other conditions that don’t often trouble terrestrial networks. DTN therefore adds some store-and-forward tech, so that packets aren’t dropped for lack of a node to contact.
The SIG’s meeting considered more internetworking ideas to advance DTN, assessing their ability to prioritise high value data, efficiency, and ability to scale to 100,000 nodes.
[2]
Candidates are detailed in this [3]summary report from the August meeting. Among the schemes considered was one called “Spray & Wait” that assumes nodes have no knowledge of the network, but in which nodes figure out the optimal number of copies of data that can be in transit. When traffic reaches a node, it is forwarded, and eventually the network figures out how many copies are bouncing around and stops routing any other than those already en route to the destination node.
[4]Bonus features: Sony uses Blu-ray tech to simulate 466Mbps laser link from the stratosphere to space
[5]South Korea says 2022 moonshot on track, will test interplanetary internet and search for water
[6]Interplanetary Internet about as useful as flying pigs says Vint Cerf
Spray and Wait was one of five approaches that reached IPNSIG’s stated goal of scaling to 100,000 nodes but did not meet other criteria.
While the SIG has called for work to begin on the routing framework, just who will do the work, select the relevant techs, or decide a development process remains unresolved. ®
Get our [7]Tech Resources
[1] https://ipnsig.org/2022/09/19/september-newsletter-2/
[2] https://pubads.g.doubleclick.net/gampad/jump?co=1&iu=/6978/reg_onprem/networks&sz=300x50%7C300x100%7C300x250%7C300x251%7C300x252%7C300x600%7C300x601&tile=2&c=2YymPOCFVzkctuCeyAhdsUgAAAME&t=ct%3Dns%26unitnum%3D2%26raptor%3Dcondor%26pos%3Dtop%26test%3D0
[3] https://docs.google.com/presentation/d/1p8pI_Hs71GhGGjQJCZDGnzb2YXKtIU1B/edit#slide=id.p9
[4] https://www.theregister.com/2022/01/28/jaxa_sony_446_mbps_space_network/
[5] https://www.theregister.com/2021/08/31/south_korea_says_2022_moonshot/
[6] https://www.theregister.com/2015/05/27/interplanetary_network_sig/
[7] https://whitepapers.theregister.com/
Only 100,000 nodes for the whole solar system? It might seem enough now, but...
Is it April 1 already?
Pointing
The thing is space is BIG, and to get any decent radio or optical link you need to focus it really tightly. And that means to sender/recipient generally has to have directive antennas pointing at each other. So it is not just down to how you bounce data around, you also need mechanisms to make sure various nodes can align themselves for comms when needed, and hours later when a reply might be due.
So they all need to be passing physical coordinates / propagation vectors as well so they have some sort of knowledge of this.
X25 anyone?
Yup. OSI (Open Space Initiative) protocols work better.
I was briefly involved with some Interplanetary Internet stuff years ago, which mainly seemed an excuse to get Ciscos in Spaaaaace. Best part was chatting with some actual NASA engineers about the protocols already in use. Oddly enough, they were already rather effective and efficient, and so I wondered why we'd want to burden them with IP.
Now we have 'Spray and Pray', and still don't seem to think that having a telecomms-style destination octet. 1.x.x.x.x(:x)= Earth, 10.x.x.. = Moon etc. But then most of the Internet's routing problems could've been improved by copying the way telecomms did numbering well over a century ago.
Sounds great...
But how are you going to pwn n00bs on Counter Strike with that space lag?
UUCP still works quite nicely.
Covers all their actual needs (not necessarily stated needs), and more ... runs on minimal equipment, can use Earth-bound networking schemes for transport as required, etc. etc.
Sorted. No charge.