|
Non-scalable? How? I've had no problem scaling.
|
|
|
|
|
There is more than one way to look at scalability where the typical way is linear (Big O(n)). But this leaves out another important factor, the centralized SQL/RPC's dependence on single-enterprise industries here called server farms. There is a probability and lifetime associated with these enterprises. GE and IBM have lasted a long time but both companies are fractions of their former sizes. The same is true for server farm enterprises. Just a wild guess but I'd say there is a 50% chance that Microsoft's Asure will collapse within 10 years. Same goes for AWS, Google and Apple. I might be wrong as much as an order of magnitude, but it is indisputable that single-enterprise centralized server farms' existence is a non-negligible factor.
So when I say non-scalable, I am combining superficial linearity that the RPC/SQL algorithm suggests (O(n)), with their fallibility (fx, O_p(0.5) over fx 10 years). Then from that perspective, the current set of server farms is not scalable. And I haven't even considered other even very real factors like security with respect to facility centralization. Note that, when I say failure, I don't mean existential failure (ie, breaks down completely). I mean failure like for example, the enterprise's price keeps climbing losing more and more customers. Or attacks increase and the company fails to fix the problems in a timely manner if ever. Etc.
|
|
|
|
|
Your argument might be believable if Azure's hardware underlying the "cloud" were static. But, like any remote site, cloud or not, the hardware is upgraded as needed. That changes the calculation since newer hardware will improve in speed, reliability, and quantity, and as it has for decades, lower in capital cost.
Since you did not mention that your definition of scalability has little to do with the commonly used definition, it is natural that readers will take the usual definition.
You also overlook the competitive factor in keeping prices down. Price pressure generally results in a compensating rise in ingenuity that lowers costs.
If you do not like the cloud providers, you might find a better argument than something this.
|
|
|
|
|
There is no doubt that Big Tech will continue to shore up against competition and keep pouring resources into their model in the hopes that their "anything can fly if you give it enough power" model/belief has no limits. Noone really believes that, though. My point was to critique the assumption that nothing will happen that will upend the centralized model probably within the not-too-distant future, fx, <10 years, which, if true, puts a big hole in the linear-scalability-only prediction model.
|
|
|
|
|
I think you over-exaggerate the issue when it comes to hardware as the cause.
I do agree there might be something that causes many current customers to shift back to using their own sites, though.
Cloud vendors rely on the Internet and private telecommunication trunks being stable and nearly 24x7x365.
All it takes is a determined, organized, enemy like the ChiComs, Russia, or Iran to use hidden terrorists to covertly and physically disrupt communication nexuses. That is low tech, easy in and out, and only needs to be in a varying set of remote locations.
Inconsistent TCP/IP connections, or worse, high latency connections, would cause disruptions to business that would negatively affect revenue and/or net profit. Then they would look to bring their systems back in-house where access by the company users is a private, physically local, network.
Just a few such low tech attacks on nodes where Internet communication lines have junctions or are exposed enough, primarily rural areas, would be enough to scare enough cloud customers away and significantly drop cloud revenue below what is necessary to sustain cloud operations profit.
A wise cloud vendor should already have an internal project to package their cloud software for those wanting a private, collocated, cloud. Intelligent enough to discover what is there on the network, and prompt the installer on what they can install under the license.
But are any cloud vendors smart enough to see that market now and over the next 10 years?
|
|
|
|
|
I don't see "on-premises" as a general option for a future disgruntled enterprise customer. I know that the binary choice of one vs the other is the current way of thinking. But actually, when you think about 'cloud' and 'local' in their limiting cases (ie, all cloud vs all local), then compare the two, you only really have somewhat of a case to go on-prem if you are a huge enterprise since the other limiting case is a single human being doing computing with himself, like playing cards on a PC. But even for large enterprises, detailed connections with the outside world are just as imperative as they are for a small business or even individuals doing anything on the computer.
The above is true and will remain true, in my opinion, as long as the case for individual human beings doing computing remains impractical regarding establishing sets of durable end point to end point connections with the outside world. From this scenario the cloud is merely a kind of a superfluous man-in-the-middle. But that discussion is too far away from the OP's disaster shout-out.
|
|
|
|
|
I know I'm a bit late, but the first rule for Cloud services is: add a notification about estimated costs.
I do the same for a customer using some free Azure services with a pay-as-you-go subscription, where the only paid service are a few SQL Server.
Never paid more than 70β¬, using a few Sql server, Application Insights, storage, etc.
|
|
|
|
|
Kiss your money goodbye.
I have to tell you that I have hundreds of Azure accounts with more than one subscription, and this is the biggest item I make my techs look at.
When you add a Resource, mind what subscription you put it in.
You are required to use a CC (unless you are direct bill, CSP), but you should FIRST READ all documentation about it and how to apply your subscription.
It allows for you to have benefits from Non-Profits and complete the functionality by paying, direct or CSP1 or CSP2 lowering your cost.
At least Azure always had resources and resource groups to lump up your charges and you can add TAGS to do your reconciliation.
Not the same with Amazon where they hide items after you remove a resource, and you cannot remove them until you call them and 3 months later you get the charges dropped.
Good luck!
|
|
|
|
|
Wordle 1,144 5/6
β¬π¨π¨β¬β¬
β¬π¨π¨π¨β¬
π¨β¬π¨π¨β¬
π¨π¨β¬π¨π¨
π©π©π©π©π©
|
|
|
|
|
Wordle 1,144 2/6*
β¬π¨π¨β¬β¬
π©π©π©π©π©
|
|
|
|
|
Wordle 1,144 3/6*
π©β¬β¬β¬β¬
π©β¬β¬β¬π©
π©π©π©π©π©
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
β¬β¬π¨π¨β¬
β¬π¨π¨π©π¨
π©π©π©π©π©
In a closed society where everybody's guilty, the only crime is getting caught. In a world of thieves, the only final sin is stupidity. - Hunter S Thompson - RIP
|
|
|
|
|
Wordle 1,144 3/6
β¬β¬π©β¬β¬
π¨π¨π©β¬π©
π©π©π©π©π©
Within you lies the power for good - Use it!
|
|
|
|
|
Wordle 1.144 5/6
β¬π¨β¬π¨β¬
β¬β¬π¨β¬π©
β¬β¬β¬π¨β¬
β¬π¨π¨β¬π©
π©π©π©π©π©
I could have used my second and third usual starters and got it early. That's a tough blow.
GCS/GE d--(d) s-/+ a C+++ U+++ P-- L+@ E-- W+++ N+ o+ K- w+++ O? M-- V? PS+ PE Y+ PGP t+ 5? X R+++ tv-- b+(+++) DI+++ D++ G e++ h--- r+++ y+++* Weapons extension: ma- k++ F+2 X
The shortest horror story: On Error Resume Next
|
|
|
|
|
Wordle 1,144 4/6
β¬β¬π¨π¨β¬
π¨π¨β¬β¬β¬
π¨π©β¬β¬π©
π©π©π©π©π©
|
|
|
|
|
Wordle 1,144 4/6
🟩β¬β¬β¬β¬
🟩β¬🟨β¬β¬
🟩β¬β¬🟩🟨
🟩🟩🟩🟩🟩
Ok, I have had my coffee, so you can all come out now!
|
|
|
|
|
Wordle 1,144 4/6*
β¬β¬π¨β¬β¬
π©β¬β¬π¨β¬
π©π¨β¬β¬β¬
π©π©π©π©π©
Happiness will never come to those who fail to appreciate what they already have. -Anon
And those who were seen dancing were thought to be insane by those who could not hear the music. -Frederick Nietzsche
|
|
|
|
|
Wordle 1,144 2/6
π¨β¬π¨β¬β¬
π©π©π©π©π©
"A little time, a little trouble, your better day"
Badfinger
|
|
|
|
|
Wordle 1,144 2/6*
π©β¬β¬π¨β¬
π©π©π©π©π©
|
|
|
|
|
I have a devkit with an MIMXRT1170 on it. That's a nasty 1GHz ARM Cortex M7 based monster with a secondary 400MHz M4 core as well. It can take DDR3! it can take NVMe, i think. (at least my devkit has an M.2 slot), it can run linux even, and is the only Cortex M (real time chip) that I'm aware of that can. Usually they run like FreeRTOS, or Zephyr RTOS. The trouble with it is the tools, and the MCUxpresso development framework. Working with it is like pulling teeth so I'm really not inclined to and it has been collecting dust, despite how neat it is.
I have an ESP32-S3 with a 7" 1024x600 capacitive touch screen. I don't know what to do with it. It doesn't have much I/O - a couple of "Grove" ports
I have another one with a 4.3" 800x480 capacitive touch screen and a CAN interface which could be interesting, but I don't really want to hook it into my Toyota right now.
The problem with having built libraries for every damned thing is there is very little in the way of new frontiers - at least ones I want to traverse.
Of all of these, the MIMXRT has captured most of my interest, but coding it is work. Like, they made sure that it was a chore. Awesome hardware, but their tools leave a lot to be desired.
Check out my IoT graphics library here:
https://honeythecodewitch.com/gfx
And my IoT UI/User Experience library here:
https://honeythecodewitch.com/uix
|
|
|
|
|
honey the codewitch wrote: Awesome hardware, but their tools leave a lot to be desired.
That reminds me of my times in the plc world.
German joke: SIEMENS (Super Im Ersten Moment, Erst Nachher ScheiΓe)
Translation: SIEMENS (Super at first sight, later just crap)
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
I worked through them for a bit. Turns out they do headhunting contracting for software developers too.
Check out my IoT graphics library here:
https://honeythecodewitch.com/gfx
And my IoT UI/User Experience library here:
https://honeythecodewitch.com/uix
|
|
|
|
|
Yeah, a former co-worker of mine changed to their beta tester unit for the new versions
M.D.V. 
If something has a solution... Why do we have to worry about?. If it has no solution... For what reason do we have to worry about?
Help me to understand what I'm saying, and I'll explain it better to you
Rating helpful answers is nice, but saying thanks can be even nicer.
|
|
|
|
|
Quote: I have a devkit with an MIMXRT1170 on it.
Suggestion: Use the Arduino IDE to program it*.
BTW I am dreaming of an M4 as primary (only) core (currently I'm using an M3 for the purpose).
*First you have to port it to the Arduino world, but, hey, that's the fun.
"In testa che avete, Signor di Ceprano?"
-- Rigoletto
|
|
|
|
|
There's scuttlebutt on the PJRC forums about the next Teensy using an MIMXRT1170.
That's Arduino, so we'll see. I love the Teensy boards.
Check out my IoT graphics library here:
https://honeythecodewitch.com/gfx
And my IoT UI/User Experience library here:
https://honeythecodewitch.com/uix
|
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
