Basecamp decamps from cloud: 'Renting computers is (mostly) a bad deal' Listening to the vendors, the analysts, hell, even the government – the direction of travel is obvious. All your server-side computing resources will move to the cloud, like it or not. Bucking that inevitable trend, however, is David Heinemeier Hansson, CTO of project management platform Basecamp. "Renting computers is ( …
My partner's employer has spent the last year moving away from Adobe products entirely to free or one-off alternatives like DaVinci Resolve, GIMP, Blender etc.
It's taken a huge amount of effort, but they seem to be convincing even the Adobe fanatics that there are alternatives.
I saw a link to another article yesterday quoting something like 80% of survey respondents had their CIOs demanding they halt or reduce cloud spending. Good to see. Cloud can make good sense in some cases but it's really more on the SaaS side(and even then not all cases). IaaS has been a disaster from the beginning, PaaS not far behind for other reasons(as such PaaS has had by far the least adoption I believe among the *aaS things).
Myself I realized this situation about 12 and a half years ago shortly after starting my first investigations for cloud. My first potential cloud project had an ROI of doing things "in house" vs in cloud of 8 months, and that was assuming the cloud solution worked flawlessly which I was sure it would not(due to many factors), but it's hard to quantify that at that stage of planning.
Saved my org(which was "born" in the cloud, so one can't say they "lifted and shifted", the people doing the cloud work had at least 2+ years of prior cloud experience so they weren't newbies) well over $10M since moving out in early 2012(while providing greater performance, security, and reliability at the same time). My manager at the time projected the ROI of bringing stuff in house at 7 months. I had to re-justify data center again many times over the decade as executives came and went, some tried to push cloud again, fell on their faces every time as they faced the massive cost differences.
Offered to save my previous org several million back in 2011 moving out(tiny startup spending upwards of $500k/mo on cloud), everyone in the company was on board except for the board so I left shortly after that, that company is long since dead.
"that company is long since dead"
Where I work IT is making a mad dash for the cloud. The only thing to save us is that the stampede is industry-wide, so whether we remain competitive will not depend on the cloud-vs-on-prem column. However, it will represent a massive transfer of wealth from our shareholders to those of the cloud providers.
MBA management buying into the cloud, thinking that On Prem, then COLO is too expensive, but renting a COLO and their Hardware was going to somehow save more money in the long run. The mythical "other companies" were going to cover all that HR/bonus/profit of the host company because "our team" is such wonderful negotiators. Nope--they are complaining about the high costs of cloud and why they spent so much money for hypothetical expandability their established customer base never needed. They massively overpaid for what they got, in ways they were too blind to foresee.
I remember observing more than twenty years ago, "The answer to any interesting problem in computer science is 'it depends'." Or, as the Right Pondians put it, "horses for courses".
Our entire IT org is less than 15 people. Most are the programmers that maintain the applications that our entire company runs on. It's <10qps. Our entire cloud spend last month was about $5k. For that, we get all the love that AWS & Heroku can give a small company.
We are getting a GREAT deal.
You can save a lot of money if you don't carry insurance. You'll probably lose a lot of business as well, because clients understand the importance of limiting risk. A substantial value of the cloud is in mitigating risks--which is why I'm so strongly opposed to using Azure.
I learned SRE at Google. I know what it takes to deliver 5 nines at scale. If your business is going to take a reputational hit if your miss 4 nines, you better understand as well before taking it on yourself. What would a black swan event do on that front? Can you handle it?
If you're not in three different data centers in three different parts of the US, with onsite staff, you're not 5 nines. (I pity those in a small country with PII or PHI.) That requires tripling your hardware AND maintaining failovers, by the way. Or you could be in four DCs in four different parts of the US and only double the hardware. With staff at each location.
Of course, economics change. It should be expected that AWS & the like want to go the route of Oracle. (And by the way, just how IS Oracle's cloud business doing?) A rapidly growing company should probably evaluate it's on-prem / in cloud position every year, a stable one every two.
Beancounters and the C-suites they advise firmly believe in trading investments for rentals because it improves the balance sheets. They outsource the cloud. They outsource jobs (and not just in IT). They outsource office equipment. Some even sell their buildings and rent them back.
Penny wise but in no way pound foolish is their motto.
Yeah. They don't mention that what they are really outsourcing is blame. Something in-house goes tits-up and takes a fortnight and £100,000 to fix ? OMG! Witch-hunt and drama! Same thing going tits-up at a supplier which take six months to fix and costs you £10,000,000 in lost business? Meh! Nothing we can do, it is the fault of those guys!
(you are not anon I'm afraid to say!)
But been there mate. Internally "things take up to three days!" , so when outsourcing the the SLA has to be 3 days. Suddenly outsourced EVERYTHING takes 3 days, so they manage expectations, meanwhile after 3 days tickets are closed for absolutely no reason so stats meet expectations.
It's so transparent - I don't get it.
Also something about this article triggered me, clearly!
"Unlike CapEx, OpEx has no or low upfront costs and allows companies to spread their expenses over a period of time. Operational expenses are included in the income statement of the company for the period during which they are incurred. For tax purposes, OpEx purchases made in a single tax year can be fully deducted. There will be no amortization of these expenses since these items are fully consumed in the tax year."
so you can use all of an OpEx against profit in the current tax year but only a bit of a CapEx (amirite? I am not a beancounter)
I think the tax implications largely come out in the wash because you can discharge the capex over the deprecation period of the stuff you bought. Actual accountants may disagree with me.
My very-possibly-incorrect understanding is that, say if I spend £10/year in opex for the next 3 years, then I can use £10 of it against tax each year in the year that that £10 was spent. Whereas if I spend £30 in capex up front and then deprecate the purchase over 3 years, then I can use £10 per year as the object's nominal value falls off. e.g. by the end of year 1, the tax system views me as having spent £30 on an asset that is now worth £20, so it counts as -£10 of profit there.
OK, this makes sense.
Using more realistic figures (£12-15 a year opex vs 30 over 3 years capex) this sounds like the messed up thinking that maybe true.
I assumed having that £20 at the end of the year 1 didn't count against you, because it was spent on something with value. So you have an additional asset in your + column
And then after 3 years, even though you've written off against tax the whole upfront cost, unlike Opex, you aren't really left with nothing, you're left with servers that can continue to run your product for another 5-10 years. Blows my mind.
Feels backwards, like production lines getting rid of robots to replace them with people.
Mix of reasons, some good, some I am not convinced are good.
One of the good reasons is cash flow: capex happens up front. Another reason is time-value-of-money, which is the observation that a quid today is more useful than a quid tomorrow.
A reason that I think is less good is that many investors use metrics like ROIC which put the company's total capital on the right hand side of a ÷ sign.
This whole article is so full of straw man fallacies it's like a Wizard of Oz convention held in a cornfield in Iowa on Halloween...
WTF does the design choices Vint Cerf and Bob Kahn made at DARPA back in the late 60s/early 70s have to do with whether you host in the cloud or not? That's just some kind of weird virtue signalling bollocks. And if you're trying to tell me that your self-hosted solution has a lower risk of going down than a well-architected multi-az or multi-region solution in AWS or Azure then you're smoking the curtains.
Anyway, regarding the economics, this is the key line is - "We're paying over half a million dollars per year for database (RDS) and search (ES) services from Amazon.".
So let's say I expect to make a gross profit of $1m this year. If I take this $500K of OPEX out, I make $500K nett profit in this years P&L.
If I can CAPITALISE that $500K and straight line amortize that over 4 years (i.e. $125K/yr) and I take that depreciation charge into the P&L then $1M-$125K = $875K and hey presto, I've made an extra $375K of profit this year, albeit I now have a depreciating asset on my balance sheet. But if my balance sheet is pretty healthy that's probably not much of a concern.
And it gets even better than that... if I am reporting to investors (or my annual bonus) is calculated from EBITDA (Earnings BEFORE interest, taxation, depreciation and amortization) my profit is now $1M... and perhaps my bonus target is now met and happy days...
So IMHO I'd bet that this whole move is driven by what the CFO wants to make the P&L look better, and all the rest is post hoc rationalisations.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
