The Big Cloud Scam

NOT primary file or data storage, and definitely NOT core business functionality!

The fact is, you could find yourself on the end of a law suit if your customers could demonstrate that, given the recent facts mentioned above, you were actually negligent simply by using a particular piece of software, or perhaps even by putting it in the cloud in the first place.

All courtesy of the same people currently cuddling up to the biggest war criminal in history. It’s hard to imagine the defence in a court of law arguing that “it’s not negligence if the government stopped it being patched, even though its public knowledge… we were just doing what everyone does”.

Having your data in the cloud is not only dangerous, it’s slow. No matter how fast your connection is it will always be slower than something running locally.

Combine that with a nice healthy install of Windows, a bunch of cloud based functionality and data storage and your machine is maintaining literally hundreds of connections, most of which are not necessary… and what you get in return is a very sluggish and fragile working environment.

For example, at one place I worked at a few years ago (a huge public sector organisation), if you accidentally let your mouse hover over someones email address in Outlook it would try and pop up an info-card. Problem is it’d take between 20 and 30 seconds to load and locked up the machine for the duration.

There was similar behaviour across the board… failed data loads, minutes waiting for Excels “Power Query” to do something a little bit of VBA could do in milliseconds, 20 to 30 minute waits for as few as a million rows of data from an Oracle BI install etc., the list goes on.

And to all those Microsoft supporters getting defensive… Yes I know there are reasons why Windows is such a sloth, that doesn’t make it ok.

To cap it off the inevitable happened last year when Crowdstrike released a file full of zeros and bricked half the worlds computers. Anyone remember the phrase about having too many eggs in one basket? Well in this case it was too many eggs in a soggy paper bag.

The cause was literally nothing… A single file full of zeros that should have been rejected. And for the MS supporters out there “No… it couldn’t have happened to any operating system, only Windows is that fragile”.

However you wrap it up, having your data and functionality in the cloud is at best a stop/start experience and in all likelihood, slower that it ever was 10 years ago.

Maybe you don’t mind a few minutes here and there waiting for your computer to get round to it, but for organisations with thousands of employees it all adds up.

Perhaps the drop in performance can be compensated for somehow now we don’t have all that complicated hardware to worry about? Actually no… even though it’s never been easier to set up pretty much any kind of server and infrastructure ourselves using tools like Docker, building such infrastructure in the cloud is massively overcomplicated.

This isn’t (just) because it’s a complex subject, rather it’s because the providers of “the cloud” are so greedy they’ve broken everything down to the finest level of granularity possible, so what might have been a simple six step process back in the days of physical infrastructure is now likely to be so complicated it takes 5 times longer and is reduced to a diabolical form filling exercise. No real skills, just patience and a low expectation of life.

At one place I witnessed, over a period of nine months, an off-shore tech company try to migrate a single SQL Server (in dev, test, pre-prod and prod environments). By the time I moved on they’d been at it over a year and were having daily conference calls involving up to twenty people for up to an hour!

The team I worked with for the same client a few years earlier would have turned the job around in 24 hours. But back then they had a server room and low paid technical staff. Now they’ve replace the techies with highly paid stuffed suits and the costs are going up and up every year, while they get less and less in return for money that isn’t even technically theirs to start with.

Not even Microsoft could deliver on one project I was involved in. All we needed was some realtime file storage we could set user level permissions on. After a month of me pointing them at some MS documentation that claimed it was possible, they gave up and refunded the money.

They kept trying to add thousands to the costs, even said we’d need a local AD server (that’s right, a local AD server to support cloud storage with permissions)… eventually one of them actually confessed that Azure was still very much a “work in progress”.

That was Microsoft. I went online after the last call with them, found a company called Liquid Web and everything was up and running within an hour or two. Yay for the big boys… not.

You pay for everything… even the trackers

In simple terms: When you move to the cloud you move to a pricing model based on consumption in respect to things like size, bandwidth and processor ticks. Previously you’d typically have relatively fixed monthly costs like electricity bills and techie salaries.

Now a simple bug in your code can add two or three zeros to your bill in a month, before you wouldn’t have even noticed a change in your electricity bill.

Every tracker, advert, API call and file request is now part of the bill. All that hardware you could buy and depreciate etc. has been replaced by subscription models that creep up every year.

Paying by such units of measure is much more expensive than the electricity it takes to handle them, and for many the cost of the staff it took to maintain them.

If I look at my DNS server logs to see how much unwanted traffic is being blocked, it’s typically between 25% and 30% of all traffic. That’s 25% to 30% of the money I spend on network communication that’s of zero value to me… the guy paying the bill.

How much does that scale up to for a business?

Many businesses have either migrated straight back out of the cloud once the full costs became apparent, or are stuck because now they don’t have techies, they don’t have the resource to correct the mistake.

So what about all those 9’s…

Surely with all those resources and the economies of scale enjoyed by such providers as MS Azure and AWS we can be sure to get a more robust and reliable service. Microsoft even claim 99.999999 etc. percent up time. Well as impressive as that sounds, the fact is for every one of those nines there’s an infinite number of zeros embedded in the terms and conditions when it comes down to liability for a major outage.

Read the terms and conditions.

Like I said before, it’s never been easier for us to build our own on-premisses virtual infrastructure with tools like Docker, and most of the time thats exactly what the big providers are doing themselves.

Broken promises

Summing it all up, and there’s so much more… but of the main motivators for moving to the cloud, i.e.:

  • Reduced costs
  • Increased reliability
  • Increased security
  • Better performance

None have delivered as promised.

One Last thing (funny, I promise)