DISQUS

firebranded · 5 months ago

I was going to say that VC's who don't mind wasting money will let their startups use Amazon's or whoever's cloud. But then I gave it another thought and actually as a startup that's fine, to use the cloud to prove out what you are doing, but eventually the cost of that service is WAY more than doing it yourself. Using the cloud makes sense when you don't have IT expertise in house, but then are you really a hightech startup if you don't have that? Using the cloud makes sense to absorb extreme growth while you catch up by building it out yourself.

Give this some thought. Cloud isn't spreading the cost of one computer over many, unless you aren't using much compute power. Cloud is adding cost/profit between you and your compute power. It just doesn't make sense if you are semi stable and need full compute power. If you either aren't very stable, or you don't need a whole compute power to yourself then yeah it makes some sense.

But let's not forget Amazon and others are making a profit off of your inability to run your own Data Center.

By the way... I DON'T run a data center, nor do I want to, but I did look into moving our services off to a cloud (Amazon) and given our traffic it just didn't make sense, it would cost me more to use them monthly than to add a server a month that I am currently adding to keep up with growth. And when the growth stops (and eventually it will, even for Facebook), the costs just keep rolling in every month. With the model I looked at, when our growth stops our costs drop, when that happens our profit goes up. So why would I continue to pay my profit to another company to pad their profit? Just really doesn't make sense.

Wake up and Smell the Coffee...

Greg · 5 months ago

Thats a very simplistic way of looking at things. I do agree that as you scale the logic behind having your own infrastructure becomes more obvious. In the beginning though if you want scaleability, load balancing, redundancy, backup.. etc cloud offers that at pennies vs investing in your own infrastructure. There of course is a breaking point, how else would the cloud make any money.

The grass is always greener on the otherside. 5 engineers coding away don't need an IT guy, they're all very knowledgable with technology. However, someone within those say 5 people owns the cloud and manages it.. so saying they don't have an IT guy is stretching it.

Timmy · 5 months ago

Venture capital is the only profession where amateurs routinely outperform professionals. And that's despite the huge advantage the professionals have.

Anthony Ha · 5 months ago

Examples?

Geoff H · 5 months ago

@firebranded

The biggest advantage of the cloud is that you dont have to care if a particular machine dies. If it dies, release it and bring in a new machine with the same image.

Another secondary good is the ability to scale up for several hours, instead of several months.

This is in comparison to leasing machines that are in someone else's care. You still have full root access over the machines, and no one else can log into them.

There are advantages for buying/leasing machines and hosting them in a colo, but in both of those cases, the machines show up slower than rented machines (by hour or month), and turn over on fixes is slower, but you can vary the hardware more and have ability to add more sophisticated storage and other options.

I can and have run colocation facilities, and would make different decisions for different circumstances, but the cloud is not a losing proposition in a number of cases, and the cost is competitive with leased machines in a number of circumstances.

Theres no coffee to wake up to, there are options that work better or worse in different circumstances.

JasonBrett · 5 months ago

"Fenton said – and that’s more meaningful than an analyst’s report."

No. No it's not. Anecdotal evidence does not make a trend or a market. This is an amateurish mistake that so many entrepreneurs make it's not even funny. Anecdotes are for examples, not for proofs.

That said, cloud makes sense in some cases, and not in others. The value of cloud computing is not a debate. It's a numbers and risk game. I would expect a publication catering to VCs and entrepreneurs to take a more practical approach to the question than this.

Anthony Ha · 5 months ago

I'm not sure what your objection to this post is. I think it's pretty clear that I'm not trying to offer definitive proof about how cloud computing used, just sharing two interesting anecdotes from a panel. Would you rather we just didn't cover these events at all?

In that same vein, I think you're misreading my last two paragraphs. I'm not saying that anecdotal evidence offered by a VC means more to me than an analyst's report -- that's Fenton's assertion, which he elaborates on in the final quote.

Tathagata C · 5 months ago

Yes it does make sense for startups to use cloud computing. The upfront costs are less. But as the startup start growing (in terms of everything) two things happen, both of which are psychological:
a) they have more money and they want more control
b) they start thinking (somewhat unreasonably maybe) that they can do better

Few startups make it so big that it becomes cheaper to host on your own servers.