I’m not an expert here, but I’ll wade in with some thoughts:
1.) Edge computing will come from those companies with compute servers that already handle intensive computation workloads (i.e. microdata centers).
Edge computing is more than just having computers close to the endpoints. There’s a tricky combination of what you do at the edge and what you still need from the central computer/database. This is where CDNs have some advantages, as they already have figured out how to balance what to store close to the endpoint and when to go back to the central server. Now instead of applying that to images, they’re applying it to compute cycles. As most compute requests will be unique, the legacy method of simply caching something requested once won’t work, so the more modern CDN architectures have advantages here.
2.) FSLY makes CDN tasks better/easier as FSLY moves into low/no code.
Not sure I see see Fastly as a low/no code provider. Fastly was built by developers for developers. Documentation, sample code, etc. are all online. I believe Fastly uses best of breed tools, but I haven’t seen any push by them to use 4th generation or other low code services. Maybe this is indeed something they’re working on, so that common use cases can be implemented by customers more easily. I don’t know how that would adversely affect Fastly’s ability to provide edge compute services
3.) Skeptical that FSLY is improving CDNs to cache and deliver content closer/faster to the edge.
This sounds like someone who is believing Akamai and other legacy CDN vendors in the out-dated belief that more POPs (Points of Presence) is the only/best way to deliver content faster. I have discussed Fastly’s approach in other posts here, but essentially, Fastly has their own internet backbone, built on top of Arista and Fastly owns their own SDN code.
4.) A FSLY use case such as Shopify is not a strong use case since e-commerce is not computationally intensive compared to running an autonomous vehicle, which is what edge computing will deliver.
Well, the NY Times is a better use case. Fastly enables subscribers to log-in on one of Fastly’s edge servers without having to tap the NY Times’ own central computer every time. Legacy CDNs depending on a hit cache can’t do this because a) still have to hit the central server the first time anyone logs on, and b) how does the edge know the cache is still valid (that is, whether the user’s account is still active, or was canceled, for instance)? Fastly has lightning quick updates from the central server to all its POPs, so conceivably a user logging into via Fastly does not ever create a hit to the central server.
But, yes, Fastly is new to edge computing. They’re taking their time with Compute@Edge to be sure to get it right - it’s been in Beta for months now. I’m still getting up to speed on it, but if it’s as solid and revolutionary as their CDN offering it could be a game changer.
Finally, I highly doubt that edge computing is going to deliver autonomous driving. We’ve already seen automakers from Tesla to Mercedes going with redundant high-power chips designed to handle neural networks and I doubt that even with 5G anyone would want to risk not just the latency but also the potential network downtime in such a critical situation. Remember that the edge can be the end device itself - edge is not limited to a cell phone tower or micro-data center.
4.) Not convinced that CDN will be a leader fulfilling this need, in fact, real edge computing players could make CDNs obsolete as once compute servers are at the edge, CDNs may not be needed any longer.
That makes no sense to me at all. Spotify, the NY Times, and thousands of web sites will still want the speed optimizations that CDNs provide.
I would like to know the companies Beth thinks are “rea edge computing players,” though. This link, https://stlpartners.com/edge-computing/edge-computing-compan… , covers 15 players - but they’re all small.
5.) FSLY is currently benefiting from surge in e-commerce and content consumption/streaming. And there is not much edge computing going on beyond just faster content streaming and/or lower latency for content.
Is she really saying that Fastly is doing well in a growing market is bad?