I think of the brain as a computational device: It has a bunch of little components that perform calculations on some small aspect of the problem, and another part of the brain has to stitch it all together, like a tapestry or a quilt.

Daniel Levitin

The Starlink constellation is growing incredibly fast. It is said to have already launched over 1500 satellites into a low earth orbit. The approved plan is for 12,000 satellites and SpaceX, the parent company of Starlink, has requested to be able to launch an additional 30,000 satellites for a staggering 42,000 satellites in orbit around the earth.

The general assumption is that all of these satellites will provide internet connectivity services. And, that may actually be true. But, what else can these in-orbit birds do beside turn around signals from the earth back down to end users? What else is possible?

If you consider that each satellite is just a node in a network fabric, and that each of these nodes can be virtualized, to be made to be anything and deliver any digitized features, then the expansion of the functionality of this satellite fabric is all rather exciting to ponder.

Let us consider that not all of the satellites are relay nodes. What if some of these satellite nodes provided other functions. What might these functions be? What if these satellites were stitched into a federation of nodes? If they were woven into a fabric with intricate details and quilted together. Just as a patchwork quilt is a blanket in which the top layer may consist of pieces of fabric sewn together to form a design. Originally, this was to make full use of leftover scraps of fabric, but now fabric is often bought specially for a specific design. The Starlink constellation is a space quilt sewn together to link the various disparate pieces into a cohesive and specific work of art – a space fabric network.


Of course, the primary function of a Starlink satellite is to provide communications. The network fabric is the primary need. But, are communications limited to links from earth stations to end user terminals? We already know that Starlink is planning some pretty capable laser connections between the birds, These laser connections will allow the consumers to receive high speed internet in the hundreds, thousands, or more gigabits per second data rates. That alone is a powerful capability.

These announced laser connections between satellites will result in high speed links that will facilitate better hand-offs between satellites and a more contiguous connection for the end user. The instability of these connections has been seen to be a challenge for the consistency and stability of the Starlink service during the beta testing of the constellation. Once these laser links are fully deployed and the density of the constellation increases, so should the availability and reliability of the consumer connectivity solution.

But, what else can these laser links do for Starlink?

Well, if a satellite node is more than just a basic communications node, then these laser links can become powerful input / output conduits to interconnect other functional nodes and string together comprehensive arrays of satellites into a denser space patchwork of intricate patterns.

What if they could use laser links to communicate upwards to much larger and more powerful geostationary satellites. These GEO birds can become super nodes with far greater functionality to collaborate with the constellation of LEO birds in a parent-child relationship, or a client/server structure, as we call it in internetworking.

So, communications are not just downwards, they can be side-to-side and upwards-downwards too.


If a number of the satellites in the constellation are not all meant for consumer connectivity, then what can they be used for? What if they are simply a battery of intense, super dense CPUs, GPUs, and co-processors. Each of these nodes can be configured as compute nodes. These compute nodes can run some pretty serious math in a hyper cooled environment of space. The concern with this idea is the amount of power needed to run these computations, how can Starlink power these math birds? Solar power will only get you so far. They can work to reduce the CPU / GPU circuit board density to use less power, so that is still a good thing. They have an abundance of cooling and heat dissipation in space, so that is valuable too. But, they will need a special kind of power source on these LEO birds to energize the demands of massive computational calculations. Is a baby nuclear generator practical? Whatever the power source is, it must be compact, ultra safe, and offer an abundance of energy.


If we have compute nodes, we also need storage nodes. We need LEO satellites that are adjacent to the compute nodes and able to hold data temporarily to drive calculation results and inputs in and out of the compute nodes. These RAM birds will not likely use any spinning media due to reliability and station-keeping needs and will all be solid state. Again, cooling the RAM should not be a problem as space can be used effectively for these via radiating fins. I do not envision these storage nodes being used as long-term storage, just short-term while calculations are being run in the adjacent compute nodes. Powering storage is far less of a concern then the compute nodes, but still an issue to be resolved.


We can see other satellites used to run specific programs, as an example to run Artificial Intelligence functions. The form of AI used in space will likely be deep learning as the Starlink constellation is already by default configured as a neural network. So, the AI can be spread-out over a variety of AI nodes, linked with laser, and intertwined with compute and storage nodes.


Back on earth, there will be major data centres that function as cloud nodes that feed the earth stations and provide hosting services and a myriad of other traditional cloud functions including hosting of many powerful applications, just as we do today. One of the biggest headaches of the multi-cloud model is the vagaries and varieties of the connections between cloud locations demanded by the different cloud operators. One site accepts a dedicated Gig-E, while the next site dictates a VPN over the public internet.

However, when all earth based clouds are connected with Starlink connections, these vagaries and varieties become harmonized and simplified to a common standard. That idea is very appealing.

The latency over these Starlink links are all standardized too. They are predictable. Which we cannot say when we consider the issues in predictability for various interwoven terrestrial links. If we do use terrestrial links in this mix, then they will all need to be engineered, dedicated connections over glass that provide known performance parameters.


Now, if we look at the end user terminals. today they are just a smart antenna with with a box filled with de-modulators, down-converters, and an I/O card. What if we added edge computing to the end user terminals and loaded in some basic compute, storage, and AI in a digitally connected virtualized side-car box. The AI would likely be Machine Learning due to the computational limitations of the box itself, but still, that would deliver some incredible capabilities at the edge. Especially when the edge is coupled to the sky and back to the earth with all of the other nodes and clouds.


Does SpaceX truly need 42,000 communicating nodes to provide comprehensive coverage to the world? Will every satellite just deliver connectivity as the early deployments are doing today?

So, if this daydream comes true, what applications can be run on it? How can it all be used? Will it be abused? Will these orbital shells wrapping the globe be used for us, or against us? While the positive side perspective is both exciting and compelling, the negative side point of view is rather unnerving to consider.

I find it hard to comprehend that 42,000 LEO satellites are essential for communications only. That is a whole lot of satellites. So, why so many? What types will they be? What functionality will they bring to the space fabric? Only time will tell….as Elon Musk is not sharing his vision with us yet….

————————–MJM ————————–

About the Author:

Michael Martin is the Vice President of Technology with Metercor Inc., a Smart Meter, IoT, and Smart City systems integrator based in Canada. He has more than 35 years of experience in systems design for applications that use broadband networks, optical fibre, wireless, and digital communications technologies. He is a business and technology consultant. He was senior executive consultant for 15 years with IBM, where he has worked in the GBS Global Center of Competency for Energy and Utilities and the GTS Global Center of Excellence for Energy and Utilities. He is a founding partner and President of MICAN Communications and before that was President of Comlink Systems Limited and Ensat Broadcast Services, Inc., both divisions of Cygnal Technologies Corporation (CYN: TSX). Martin currently serves on the Board of Directors for TeraGo Inc (TGO: TSX) and previously served on the Board of Directors for Avante Logixx Inc. (XX: TSX.V).  He has served as a Member, SCC ISO-IEC JTC 1/SC-41 – Internet of Things and related technologies, ISO – International Organization for Standardization, and as a member of the NIST SP 500-325 Fog Computing Conceptual Model, National Institute of Standards and Technology. He served on the Board of Governors of the University of Ontario Institute of Technology (UOIT) [now OntarioTech University] and on the Board of Advisers of five different Colleges in Ontario.  For 16 years he served on the Board of the Society of Motion Picture and Television Engineers (SMPTE), Toronto Section.  He holds three master’s degrees, in business (MBA), communication (MA), and education (MEd). As well, he has three undergraduate diplomas and five certifications in business, computer programming, internetworking, project management, media, photography, and communication technology. He has earned 20 badges in next generation MOOC continuous education in IoT, Cloud, AI and Cognitive systems, Blockchain, Agile, Big Data, Design Thinking, Security, and more.