$NET Piper Sandler Downgrade: My Thoughts

Investment Analysis Clubs Saul’s Investing Discussions
1

This Piper Sandler downgrade took place after the close. I have never read a “rave” downgrade like this. Very positive. Just saying the price has gotten ahead of itself, although they raised Revenue guidance by 2-3%.

https://twitter.com/StreetGuruHQ/status/1448011465164824579

As I stated on the Destiny Solutions board, I did not sell a share on this news, as did many traders in the after-hours yesterday.

This Downgrade might even pick up the stock and turn it green because Piper Sandler is so bullish on this company’s future, just not high on the current Market Cap. (This is happening today.)

I would welcome a 20% or more downdraft in $NET and then look for another buy opp. And if we set another All-Time High today, I’m good there as I bought more $NET and added to my position this past Monday. Machts nicht.

I wish to do some spitballing on edge computing which this Downgrade never mentions in an area this analyst neglects when talking about prospects for $NET:

Where I believe mega-growth will happen (and I’ve yet to see an analyst start thinking about this) will be in the future with TaaS (or Transportation as a Service), and by “future,” I mean during the next 10-years. Think Netflix’s growth in share price when it was switching from mailers of DVDs to streaming, and streaming in the early days was janky, latency-ridden, but easier than waiting on the mailman.

I submit that $NET is building a “new” type of cloud-focused Edge Computing to propel the IoT in cars. More importantly, $NET will be mission-critical for one of the most important future markets: autonomous vehicles.

Autonomous vehicles will need three primary things for genuine Level 4 and 5 computing to happen:

  1. An unassailable big brain A/I software/hardware embedded in the car
  2. Blazing speeds from the Edge Computing
  3. And un-hackable security

The Big 3 cloud computing names, $GOOGL, $MSFT, and $AMZN, are already moving into this same TaaS market with edge computing, but these three are not auto “agnostic.”

$GOOGL is known for its sibling Waymo, which, if you believe scientists from MIT, Stanford, Oxford, etc., is the best “geo-fenced” offering on the planet when it comes to autonomous driving.

$AMZN bought AI autonomous driving company Zoox, and they and $F invested heavily in Rivian.

$ MSFT’s Azure Edge Computing has recently partnered with $GM and their autonomous driving AI offering, Cruze, which they hope will be as good as Waymo for future ride-sharing.

So where does this leave $NET?

$NET would be a perfect match for any Autonomous AI being released to the market by companies who don’t want their drivers, driverless cars, or A/I “owned” by one of the Big 3 data sellers. That’s one way. (There will likely be a fourth “BIG” if and when $APPL releases autonomous software and the rumored $APPL car. And we know Apple isn’t going to share its A/I OI and their own Edge Computing with the other three if and when this happens, nor would the Big 3 want to use Apple software.)

But here’s where the opportunity arrives for $NET when it comes to TaaS:

Every car manufacturer or A/I Autonomous Drive software builder bringing entertainment, gaming, concert tickets, table-bookings at restaurants, OTA software updates, web surfing, weather, traffic reports, GPS ad suggestions, and online shopping to cars in the future will need speed and security as their paramount concerns.

(Don’t suggest to me stuff like Tesla Starlink satellites or the similar sat system Amazon is constructing. If homes on Earth can’t get reception on rainy days, I can’t entertain any fanboi nonsense which doesn’t make the jump to fast-moving cars. Furthermore, there are bandwidth limitations for sat links which prevent them from being used in Edge Computing. As it stands right now, latency issues and frequent blackouts with Starlink are mentioned often on social media by first-adopters. If swaying trees block your dish antenna in a breeze, you’re not going to get uninterrupted internet, and you will either move the dish or cut down the trees.)

Latencies and hacks can crash and burn tens of thousands of cars in one fell swoop. That’s what I am focused on going forward. And if you’ve been keeping up with the news on $NET here on Saul’s board, you know these guys are building better, faster Edge computing with redundancies.

I submit $NET is lowballing the price of its R2 cloud offering to steal market share from the “BIG 3” Cloud providers. I can also make a case that $NET might keep the R2 price low and invent some revenue sharing plan on the monthly subscriptions AI firms and car manufacturers plan to use.

Many of these vehicle companies will not use Waymo, Cruze, or Zoox, where their competitors know their data.

So this is a natural fit for our “agnostic” Edge Computing leader, Cloudflare.

Topline growth might suffer for a while, maybe years, while we quickly expand into car company after car company or supply uninterrupted data from providers of online shopping, movies, weather, news, etc.

The Piper Sandler analyst with the Downgrade here thinks $NET will turn its first profit in Q1 2023, while I’m thinking, “Profits? Who cares now during the first inning of TaaS and IoT? Invest as Bezos did in the early days of $AMZN.”

Once A/I, car companies, fintech, blockchain smart-contract providers, etc., have switched to Cloudflare for a “tryout,” Cloudflare will not have any problem keeping them as customers as long as they continue to perform and upgrade their Edge Computing offerings. Many thousands of businesses will need and want top-notch, agnostic speed and security for edge applications. (Blockchain apps are springing up everywhere in real life, and that’s another subject for another time when it comes to Edge Computing. I believe $NET will be the top-choice for independents using blockchain in gambling, NFTs, music and video sales, etc.)

And by remaining “agnostic,” it also means anything $NET keeps on the edge for others can mesh with the Big 3 autonomous driving software. Meaning, “Hey, we at Cloudflare don’t want to hold you back, but speed you up.”

Get new firms using $NET offerings in TaaS, and pricing can be fungible.

I also believe $NET is thinking ahead and will win the day for outlier auto-manufacturers not attached to the Big 3 Cloud Computing firms.

I think of auto-manufacturers in EVs and Hydrogen Fuel Cells such as Lucid, Polestar, Mercedes, Nio, Toyota, Hyundai, etc. Their Autonomous Driving initiatives will never happen in America, Asia, and Europe without speed and security. In a world of driverless vehicles, you don’t buy second-tier edge computing because of price.

Here is a direct link to the Piper Sandler downgrade link without having to open that first link on Twitter. See how this analyst raves about $ NET’s prospects while giving it a Downgrade. ($NET is busy setting another All-Time High this morning, Wed, 13 NOV 21. So, I won’t be getting my 20% downdraft today.)

https://pbs.twimg.com/media/FBhewNEXEAU3WO3?format=jpg&n…

p.s. Talking about “agnostic” edge computing for other automakers, let’s look at Lucid’s latest news yesterday.

You probably know Peter Rawlinson, CEO/Founder of $LCID, was the designer of the very successful Tesla Model S. Three weeks ago, he allowed the press and the public to tour the first $LCID factory in Arizona. Every car magazine writer who drove one of the Lucid Dream demos raved about the precision work and performance of these cars built with painstaking manufacturing Charles Deming would have fainted over if he were alive to see it in motion. One car mag said, “The Lucid Dream is NOT the Car of the Future. It Is The Future of Cars.”

$LCID begins deliveries of their futuristic Lucid Dream cars at the end of this month, October 2021.

To give you an idea about how much compute power Lucid $LDIC needs and is offering in their soon-to-ship Lucid Dream cars as they move down highways and byways, read the following.

https://insideevs.com/news/540079/lucid-introduces-dreamdriv…

DreamDrive has a total of 32 on-board sensors, cameras, and radars, including the first automotive LIDAR in North America, as well as a multi-faceted driver-monitoring system and “lightning-quick on-board ethernet networking.” The backbone of the powerful computer network in the vehicle is the Ethernet Ring, a high-speed data network enabling four computer gateways—one at each corner of the car—to communicate with each other at gigabit speeds.

This hardware powers more than 30 features via a user-friendly interface, including collision avoidance, Highway Assist, Traffic Jam Assist, and Auto Park. The most advanced iteration of Lucid’s ADAS will be marketed as DreamDrive Pro. It will offer additional computing and sensor hardware in preparation for new features to be delivered via seamless over-the-air (OTA) software updates.

Okay, most of the above features are handled in the car, as you have just read. But think about the need for info on closed bridges, accidents, traffic jams, bad weather, available charging networks, sports betting during a game you are watching on a car’s big screen, lottery ticket purchases built on blockchain, etc. All of this happens as you move up a highway at 70 MPH. The information changes will need to be immediate and reliable as you speak to your intelligent AI built into your car’s hardware and software.

(I’ve seen demonstration videos from $LCID for the new DreamDrive Pro, and let me say, there’s nothing on the market from anybody right now which has the features which act with such excellence in parking, heavy traffic, etc., without geo-fencing. Please give us a Lucid Robotaxi in 10 years with no driver. Oh yes, the riders will be gaming, listening to music, watching the news, betting on a football game they are watching live, boogying to music videos and live DJ sets, PPV fights, whatever. The future of cars must deliver this stuff like sitting at your desk at home with an excellent net connection. If your power goes out at home, you lose connection. But there must be redundancy for Edge computing in a car, and $NET’s buildout is moving at a clip to make this happen.)

So, unless someone else decides to enter the cloud computing fray as $NET has done, I don’t worry about this stock growing the topline more slowly going forward while expanding rapidly into the minds of businesses needing Edge computing speed and security. I am thinking about what the TaaS requirements will be worldwide, and what the rewards will be like down the road. And that’s a huge, expanding TAM for $NET ten years from now.

Keep your radar tuned to when we begin hearing about TaaS on $NET conference calls. IoT will be massive, but TaaS will erect as big a money-printing business - possibly more significant - for vehicles.

You don’t want to be left behind when these mentions begin. Other car manufacturers are not wanting $AMZN $GOOGL $MSFT data collectors as gateways to their customers. Plus, A/I purveyors other than Zoox, Cruze, and Waymo, will naturally adopt the best of class agnostic Edge computing offerings from Cloudflare if, indeed, $NET still leads in this new area ten years from now.

With planning, or by helping purveyors of info for any commerce which makes $$$ off every click, I can see where $NET edge computing might work deals of revenue sharing for every click to keep things running smoothly and securely in future cars.

If not $NET, who else? That’s the job of this board: let us know about agnostic competitors.

If someone else develops an agnostic Edge Computing suite as $NET is now doing, what prevents us from investing in it too? TaaS will be an industry worth Trillions in revenues. (And I’m not even mentioning its use in planes, drones, and boats.)

Okay, that’s my wide-angle thinking about one of my Top 5 holdings and why I did not sell on the Piper Sandler news after the market closed yesterday.

Again, for TaaS - and IoT - we are in what I consider the first inning, and we keep running up the score.

$NET can continue to invest in expansion and show no profits in 2023 as the Piper Sandler analyst projects it will, and I don’t care.

The imperative here is to build “best of” edge computing that is agnostic and will have to mesh with everything out there with unmatched speed and security in moving vehicles of the future. The best part of all this is “cyber-security” starts - but doesn’t end - with $NET for Edge Computing. (Someone else did a bang up job on this board about $CRWD and $ZS joining forces with other cloud providers.)

What we have to insure is vehicles are un-hackable going forward. Otherwise, we will face a 9/11 on our highways at one specific time in a dystopian future.

103 Likes
2

And just 27 minutes ago, muji on Twitter posted this $MSFT Azure news:

https://twitter.com/hhhypergrowth/status/1448364807493652482…

muji @ hhhypergrowth

@hhhypergrowth

Global outage of all Windows VMs in Azure for 8 hours last night

The direct link to the story he references:

https://www.zdnet.com/article/global-azure-outage-knocked-ou…

Headline: Global Azure outage knocked out virtual machines, other VM-dependent services

Subheadline: A nearly eight-hour outage affected Azure users globally who were using Windows VMs and services dependent on them.

Now imagine if this happened in a GM Robotaxi (as I said in the OP, $GM and $MSFT Azure have partnered on Cruze, $GM’s autonomous driving A/I effort), and imagine if this future is where there are no drivers, no steering wheels, no brake pedals, and an onboard system which can’t retrieve important information from the Azure Edge Cloud.

Or imagine this happened and your IoT home security didn’t work properly for 8-hours and there was a break-in or worse in your home.

This is the kind of brand destruction we don’t want with $NET.

Profits? Not now or next year or the year after. Don’t bother us. We have to expand and progress.

Hire the best minds, eat market share, quickly iron out kinks, continue with redundancy, and get this thing right without hacks and long downtimes in Inning #1 while offering Best In Class security.

Do this and we will be Edge Computing Gold and no part of our business will ever suffer an eight-hour outage.

Eight-hours?

Microsoft Azure?

6 Likes
3

More importantly, $NET will be mission-critical for one of the most important future markets: autonomous vehicles.

That seems very unlikely to me. As I posted here recently (https://discussion.fool.com/what-cloudflare-is-offering-might-be… ), there are no planned production autonomous driving vehicles that require off-vehicle computing to drive themselves. In particular, since you mentioned Waymo, they don’t even keep a continuous connection to the internet (reference links in my earlier post).

Okay, most of the above features are handled in the car, as you have just read.

Actually, all autonomous driving tasks are handled in the car. Even for Lucid, although their system is far from complete and all videos I’ve been able to find are mostly CGI, so I’m not impressed. At least not yet.

But think about the need for info on closed bridges, accidents, traffic jams, bad weather, available charging networks, sports betting during a game you are watching on a car’s big screen, lottery ticket purchases built on blockchain, etc. All of this happens as you move up a highway at 70 MPH.

The information you describe as “closed bridges, accidents, traffic jams, bad weather” comes in periodically, often at the start of a navigation session (to help plan the route). But, this is not computing, it is information providing. Yes, already in many vehicles today, navigation planning is done off-vehicle, even for changes to navigation due to the driver not following the route or traffic patterns on the route changing. Indeed, this is more than 7 year old technology and does not require Edge Computing, as the latency requirements are measured in seconds, not milliseconds. And since connectivity is not available everywhere, even off-vehicle navigation efforts fall back to on-vehicle when necessary. The interesting thing is that today, with computers in vehicles getting more powerful, navigation planning is switching back to primarily in-vehicle compute (eg, Tesla).

…riders will be gaming, listening to music, watching the news, betting on a football game they are watching live, boogying to music videos and live DJ sets, PPV fights, whatever. The future of cars must deliver this stuff like sitting at your desk at home with an excellent net connection. If your power goes out at home, you lose connection. But there must be redundancy for Edge computing in a car…

Of these potential robotaxi use cases, only gaming would materially benefit from Edge Computing (because of latency requirements), and that’s hardly a mission critical use case that requires high availability, much less hardware/software redundancy.

As it stands right now, latency issues and frequent blackouts with Starlink are mentioned often on social media by first-adopters. If swaying trees block your dish antenna in a breeze, you’re not going to get uninterrupted internet, and you will either move the dish or cut down the trees.)

Antenna blockage issues are also a serious problem for 5G networks, especially from moving vehicles, with the additional complexity that the range for 5G broadcast is very short (on the order of 1000 feet) and as such it is going to be impractical to blanket every road in the US, much less the world, with Edge Computing based on 5G coverage. Good thing there’s no need. And even if you’re going to say that LTE (4G) is good enough for autonomous driving needs, well we all know about existing dead spots, not to mention whole areas of the California coast that have no coverage for miles and miles. BTW, if you’re thinking about Edge Compute enabled gaming from a moving vehicle, then you have to additionally handle the hand-off from one edge compute POP to another - a problem not yet solved to my knowledge.

The term “Edge Computing” does not demand or even imply a singular architecture, because “edge” can mean anything from a regional server to a CDN POP to on-board compute. Yes, the end-device is often thought of as “the edge,” and so are any of the hundreds/thousands of local POPs that some CDNs deploy. Moving compute closer to “the edge” is often done to reduce latency and/or save bandwidth, but it can also be to reduce dependencies on the broader internet reliability.

In the autonomous driving examples, all companies have decided to move compute all the way to the outermost edge, which is the vehicle, the end device itself. This gives the lowest latency and removes almost all dependencies on connection robustness.

Using a local POP for Edge Compute can make sense if your end device is too small, too inexpensive, or underpowered (think battery) for serious compute, or if the compute requires inputs from a number of such small devices. This is the IoT use case. But, there needs to be some additional characteristic to choose Edge Compute over the cheaper, simpler, and more ubiquitous central compute solution. Again, those characteristics could involve low latency, isolation from broader internet issues, etc. Most applications will not have those characteristics. Just because an application benefits from connectivity does not mean it meaningfully benefits from Edge Computing.

This isn’t to say that Edge Computing is a bad business, or that NET isn’t positioned well. It does say that some portion of edge computing will continue to happen on the end device, out of CloudFlare’s purview, and unless costs and ease of use change dramatically that Cloudflare’s POP-based Edge Computing architecture won’t replace most central server use case anytime soon. And it most certainly does say that there’s no reason to believe Cloudflare will materially benefit from autonomous driving or most TaaS use cases.

72 Likes
4

In the autonomous driving examples, all companies have decided to move compute all the way to the outermost edge, which is the vehicle, the end device itself. This gives the lowest latency and removes almost all dependencies on connection robustness.

Thanks, Smorgsabord, for correcting my spitball with a 180 degree curve. I am happy to learn from a more knowledgeable person such as yourself.

Question: when the Edge computing does play a part in online websites to provide IoT or push shopping, videos, photos, music, etc. into a car’s entertainment system, do you believe $NET cannot make a buck or two there?

What do you project as the big winner for security and speeding up delivery of this type information into cars? For instance, $TSLA says its on board computers are constantly feeding information back to their own in-house cloud to make their A/I smarter. Do you feel this will happen with other users of on-board A/I Autonomous Driving? That is, do you think the easiest thing for A/I Autonomous Driving software is to build their own “clouds” for collecting information to make their A/I smarter? And if $NET is in the cloud business now, would it not follow that $NET stands to benefit from what I am calling its agnosticism of not collecting someone’s raw data and selling it to the highest bidder?

Maybe I am asking the incorrect questions.

Do you think some winners will be wireless carriers in TaaS? If wireless carriers deliver content from websites whose security is protected by Cloudflare, could this not be an extension of Cloudflare’s TAM?

And could $NET’s Edge Computing in its own cloud not play a part in new auto entertainment - and maybe OTAs - if websites are speeding up delivery because of $NETs offerings?

Please continue to expound on TaaS and who might be winners in securing these different cloud businesses. As we saw today, Microsoft’s Azure was down eight hours in part of their business. Why wouldn’t this help a car maker such as Lucid use $NET instead of $MSFT Azure which is in bed with $GM?

Again, just thinking outside the box and willing to look like a fool by asking questions.

Thanks again, Beratna!

16 Likes
5

I’m not Smorgsabord, but I do have fairly deep experience in this space.

when the Edge computing does play a part in online websites to provide IoT or push shopping, videos, photos, music, etc. into a car’s entertainment system, do you believe $NET cannot make a buck or two there?

Yes, this would be similar to $NETs current CDN offerings, I don’t think of this as any revolutionary need driven by Autonomous vehicles, but as more devices coming online that need CDNs for their user experience.

For instance, $TSLA says its on board computers are constantly feeding information back to their own in-house cloud to make their A/I smarter. Do you feel this will happen with other users of on-board A/I Autonomous Driving?

Almost certainly. The only way for self driving vehicles to get better is more training data, especially around edge cases, which is most easily collected by having real people driving and then upload their data. However that data doesn’t need to get into the cloud ASAP, because it’s just going to be used for training new ML models, which aren’t going to be trained and pushed to the vehicles while they’re driving. That’s why they currently upload after the drive, I believe while connected to the drivers wifi network.

That is, do you think the easiest thing for A/I Autonomous Driving software is to build their own “clouds” for collecting information to make their A/I smarter?

Building your own “cloud” is effectively just building your own datacenter. The market is largely shifting away from that unless you’re a tech company with incredibly high compute requirements, i.e. Facebook. Most others are leveraging the public cloud providers (AWS, Azure, GCP) for all the benefits they offer over having to build your own cloud.

Why wouldn’t this help a car maker such as Lucid use $NET instead of $MSFT Azure which is in bed with $GM?

One thing to keep in mind is that $NET doesn’t currently compete with Azure, AWS, or GCP on the full catalog of offerings. There are things you just can’t get from $NET that you can from the other cloud providers. Managed databases is one example, but there are dozens.

When it comes to the Autonomous vehicle’s identity as an internet connected device, I think Cloudflare’s services provide value, similar to the way they provide value to other internet connected devices. When it comes to security, I think it’s a more complex problem than Cloudflare’s current offering would meet. Like Smorgasborg said, most autonomous vehicles will be doing all their computing on device, and as of now I don’t believe Cloudflare offers any type of cyber security offering for those types of workloads.

15 Likes
6

I’m not Smorgsabord, but I do have fairly deep experience in this space.

Thanks for weighing in!

when the Edge computing does play a part in online websites to provide IoT or push shopping, videos, photos, music, etc. into a car’s entertainment system, do you believe $NET cannot make a buck or two there?

Yes, this would be similar to $NETs current CDN offerings…

I just want to point out that the OP may be conflating CDN and Edge Computing use cases. “video, photos, music, etc.” are, as @mandeville135 points out, standard CDN use cases - simply providing content from local servers to users. Shopping requires compute, and if it’s being providing from a local server as part of a distributed network, then it is an Edge Compute use case. “IoT” doesn’t in and of itself describe what’s going on. Are you simply gathering information from devices, or gathering information from a variety of sources, figuring something out, and then telling the IoT devices to do something? The latter could be Edge Compute if done on an Edge Computing Network.

In thinking about Edge Computing for automotive and TaaS use cases, there certainly are applications, for instance uploading of information from vehicles to the OEM for later processing. This could be for problem debugging, such as analyzing vehicle DTC (Diagnostic Trouble Codes), or uploading of video streams to look at what happened just before an accident. These could all be done on a central server, but depending on the number of vehicles on the road and data privacy requirements, doing these things on a distributed network of Edge Computers could make sense.

The thing about Edge Computing, at least today, is that you need a compelling reason to go down that path. It’s just so much easier and cheaper to stand up an EC2 instance on AWS, and only if that doesn’t meet your performance targets would you start to look at something else. Edge Computing typically requires serverless APIs, which is a misnomer because there certainly are servers involved. What’s really going on with serverless is that you don’t get to store user state on the server. That’s because every time you make an API call, you could end up with a different server responding than the last time. It’s great for load distribution and having servers closer to user to reduce latency, but more difficult to program for many use cases.

Anyway, while there are many possibilities for CDN and even some for Edge Computing in Automotive/TaaS, I agree with @mandeville135 that these types of applications will not be a big driver (“mission critical” as the OP’s characterized) of future Cloudflare business. Some business, sure, but not major.

@PeregrineTrader wrote:
And if $NET is in the cloud business now, would it not follow that $NET stands to benefit from what I am calling its agnosticism of not collecting someone’s raw data and selling it to the highest bidder?

I’m not sure what this means. All the cloud providers have strong privacy controls and tenant protection from other tenants. Cloudflare may argue that their security is stronger, but that remains to be proven. I can’t recall any tenant to tenant information breach in AWS or Azure in the last decade or more.

But, the “selling to the highest bidder” is actually a Snowflake feature! Snowflake enables companies to make subsets of their data available to other companies for a fee - Snowflake provides the capabilities and acts as a broker. It’s a very cool feature that I think is going to be a driver of future business for Snowflake. But, it’s with consent only. The automotive example that floats around in my head is OEMs making driver safety and vehicle crash avoidance data available to insurance companies - anonymized of course.

4 Likes
7

“I just want to point out that the OP may be conflating CDN and Edge Computing use cases. “video, photos, music, etc.” are, as @mandeville135 points out, standard CDN use cases - simply providing content from local servers to users.”

Yes, this is what I did. Thanks for clearing this up.

“In thinking about Edge Computing for automotive and TaaS use cases, there certainly are applications, for instance uploading of information from vehicles to the OEM for later processing. This could be for problem debugging, such as analyzing vehicle DTC (Diagnostic Trouble Codes), or uploading of video streams to look at what happened just before an accident. These could all be done on a central server, but depending on the number of vehicles on the road and data privacy requirements, doing these things on a distributed network of Edge Computers could make sense.”

Okay, this is getting me somewhere. Would you consider such a use something $NET could monetize, say, in the vein of $AXON?

https://tinyurl.com/yxyw78ur

Axon Enterprise, Inc. is an American Scottsdale, Arizona-based company which develops technology and weapons products for military,[1] law enforcement and civilians.

Its initial product and former namesake is the Taser, a line of electroshock weapons. The company has since diversified into technology products for military and law enforcement, including a line of body cameras and Evidence.com, a cloud-based digital evidence platform. As of 2017, body cameras and associated services comprise a quarter of Axon’s overall business.[2]

p.s. Would you say $AXON’s Evidence.com cloud-based digital evidence platform is not necessarily kept in an $AXON built cloud, but, in - let me call it a “sub-cloud” or maybe an API as you mention next (until you can correctly define what I might be alluding to_ - and would you not agree that this so-called Evidence.com website has done a great job of monetizing raw data from “Edge Computing” as a new business for a company which once just sold Tasers? Or do we call this transferring of video from cops on the move to Evidence.com “Edge Computing” at all?

For instance, a rogue cop needlessly battered a 73-year old woman with dementia and his body cam from $AXON catches it all. Not only that, but, cameras catch him laughing about the old lady while she’s locked up for hours with no medical attention, bleeding, arm out of socket, and not making any sense because she is scared to death and not understanding why she can’t go home. Famous case, hard as hell to watch as the cop also pulls this poor woman’s arm out of socket during the arrest with the way he handcuffs her.

https://www.npr.org/2021/05/19/998433764/former-colorado-off…

So that video goes to Evidence.com for storage. The woman’s children sue in court and use this “evidence” against the two cops who brutalized this poor woman so badly she no longer trusts other humans - outside her family.

My question: Would factual proof of sadistic cops torturing a poor woman with dementia not need the ultimate security so that Bad Agents of the State would never be able to “memory hole” (disappear) this evidence?

Could $NET set something up in its new cloud offerings for say drone services (I imagine $AMZN will film every drone delivery going forward), insurers of cars (your truck’s cameras capture you, the sadistic idiot who coal rolls a file of bicyclists on a road’s shoulder because you think it is funny and in the process you run over several of the bicyclists), car manufacturers looking for defective systems in new EVs (did the car catch fire from rupturing the battery casing after running over road debris or was it self-combustion?), or store actual eyeball data from smart AR eyeglasses for $SNAP shopping (which is exploding revenues there), etc?

Moving on, you said this about APIs:

Edge Computing typically requires serverless APIs, which is a misnomer because there certainly are servers involved. What’s really going on with serverless is that you don’t get to store user state on the server. That’s because every time you make an API call, you could end up with a different server responding than the last time. It’s great for load distribution and having servers closer to user to reduce latency, but more difficult to program for many use cases.

Smorgasbord, could you come down to my level and put this in grade school English?

I had to look up a short definition of API "Application programming interfaces, or APIs, simplify software development and innovation by enabling applications to exchange data and functionality easily and securely.

So are you saying the actual data which a consumer looks over or in which they make a purchase would not be “stored” on these serverless APIs (which do use servers for speed, I assume?)

This too is important to me for my understanding. It helps me whittle down my thoughts on who will win in new ways with TaaS:

Anyway, while there are many possibilities for CDN and even some for Edge Computing in Automotive/TaaS, I agree with @mandeville135 that these types of applications will not be a big driver (“mission critical” as the OP’s characterized) of future Cloudflare business. Some business, sure, but not major.

Thank-you for this assessment. This will make me spitball in a different direction on Edge Computing use from $NET. I’ll have more questions, but I’ll take some time doing hard physical labor in the yard to formulate my next questions for you and Mandeville.

A request from me:

I do wonder if one of you would go back to the second post - after my OP from me -quoting that outage of the Azure. Here is the headline and subheadline from that post. Would you be able to explain, again in grade-school terminology, what virtual machines and other VM-dependent services? This is the kind of education I can’t find on Twitter.

Thanks to both of you, Beratna!

Headline: Global Azure outage knocked out virtual machines, other VM-dependent services

Subheadline: A nearly eight-hour outage affected Azure users globally who were using Windows VMs and services dependent on them.

Again, I thank anyone taking time out of their busy days who help answer hard questions, sometimes, not framed properly, but, I am persistent until I get to what I’m trying to understand better.

When I do this on Twitter, sometimes, and get an exasperated “smart guy” to spend time taking his knowledge and writing it down in terms that others can understand, it never ceases to make me smile how many people smash of the heart icon under my questions to show they appreciate I ask questions which might make me look dumb but which they want answered to. Many people are too intimidated by social mores and never think to ask questions because Twitter trolls will tear you to bits. Many never know or don’t have the heart to keep buggering on and they give up after one question. Or, and this is the case most often after reading so many thanks in my private DMs, they are afraid they would look stupid on Twitter for asking a question maybe 3 times, a little bit differently, to get to what needs to be answered.

I don’t care what people with more expertise in a field than think of me for asking questions. When I ask questions, I always think of Neil deGrasse Tyson, a most admirable man who delights breaking down hard concepts so that amateur astronomers or just curious people can understand physics and the cosmos. I look at experts who take time to explain things as heroes for expanding minds.

Smorgasboard and mandeville, your knowledge and patience here in answering my questions and correcting my outside the box thinking is gold to me. You both are good people.

You’ve given me new paints on my palette to brush some new wide-angle views of where this might go, or could go, or should go.

This belter is signing off now for some sweat time in the yard with my machete, the anvil shears, the chainsaw, and the mower while cranking “Above & Beyond” latest mix through the Audio-Technica MX-50s. (Yeah, I might look stupid without trendy earbuds, but, I hear the music as it was intended.)

For now, Taki and Oyedeng!

  • Rock
3 Likes
8

Would you consider such a use something $NET could monetize, say, in the vein of $AXON?

No. Cloudflare is an infrastructure company, not an application author/provider. Companies write applications and then have them hosted by Cloudflare. Or AWS. Or Azure. Or GCP. Or Fastly. Etc. I don’t see Cloudflare getting into specific verticals, at least not at the application layer.

So are you saying the actual data which a consumer looks over or in which they make a purchase would not be “stored” on these serverless APIs (which do use servers for speed, I assume?)

Maybe read this page, https://www.infoq.com/articles/serverless-stalled/ , and come back with questions, as it’s off-topic for investors. My point was only that not all applications are suitable for serverless APIs, which means they’re not suitable for Edge Computing, which requires them (as far as I know, but I could be wrong).

Would you be able to explain, again in grade-school terminology, what virtual machines and other VM-dependent services?

Again, this is off-topic for an investing board. My point was only that not all of Azure went down, only one service that it provides (Windows VMs), and even then it was still partially functional.

4 Likes
9

This is my first post here to this amazing board that I wish I had found long ago.

PeregrineTrader I have worked in IT software development for 30+ years. I wanted to see if I could address your question. There are several concepts tangled up in here.

Moving on, you said this about APIs:

Edge Computing typically requires serverless APIs, which is a misnomer because there certainly are servers involved. What’s really going on with serverless is that you don’t get to store user state on the server. That’s because every time you make an API call, you could end up with a different server responding than the last time. It’s great for load distribution and having servers closer to user to reduce latency, but more difficult to program for many use cases.

The whole idea of edge computing as Smorgasbord pointed out is to reduce latency. So if I am in NY and I am on a shopping page that requests an image the edge service would find the server closest to New York and request it from that server. All of these services are dynamically load-balanced so if the New York server is super busy maybe the request goes to the Boston server instead. Which is the point about never knowing which backend edge server you will be hitting. All of this is hidden from you as a user. Another thing these services do is keep frequently accessed content in memory so it can be retrieved as quickly as possible.

All of these services talk to each other using APIs or publically known pre-defined functions. You can sort of think of them as URLs.

Up until recently the code for these ran in dedicated servers. You had mentioned virtual machines earlier these are simply individual copies of an operating system like Windows or Linux running in a software container that is managed by some other software like VMWare. Each copy is given a certain amount of hardware resources like CPU & memory that it can use.

The idea of serverless is that there is not a dedicated virtual machine always running ready to service a request. Instead when a request comes in a smaller set of code that can execute the API request is started upon demand and it is shut down when it is done. So it is serverless in that a server isn’t always up and waiting around using up computer resources and electricity. On-demand is a better way to think about it.

Another aspect of edge computing is doing work where the data is as opposed to sending everything to a centralized cloud service all the time. You might want to preprocess data locally where it is coming in and then sending a summarized version to a central service. Think about a cell tower in New York City receiving location info from all the nearby cell phones as people walk and drive around. It’s a sea of data. So as there is more and more data there will be more use cases for computing on the edge

Hope that makes sense

34 Likes