Comment by nrao123 | Hacker News Reader

Comment by nrao123 | Hacker News Reader

This is a must read article for anybody Trying to understand what's going on in the hardware / server side for enterprise IT.

Cisco's bet seems to be that they can't do anything about hyper scale data centers (AWS, google, Azure, Facebook etc) since they will have the scale to design their own hardware.

So- they are going after large enterprises that want to still have their own servers. In this market they seem to be growing.

But - their strategy reminds of me Of trying to provide the most efficient newspaper printing machine just before the printed newspaper itself gets killed.

Two points supporting the newspaper analogy:

1) If the applications themselves (SalesForce, workday, Successfactors etc) are moving to the cloud & delivered from their own servers or AWS, Then won't the overall demand for enterprise hardware Itself reduce?

2) Even if companies want their own custom Computing resources (e.g. Analytics) - won't These sub-scale enterprises themselves go with The hyper scale data centers (AWS, azure etc) ? What on earth makes these enterprise companies believe That they can take Cisco hardware & believe that they Can be more cost efficient than AWS/Azure.

>'What on earth makes these enterprise companies believe That they can take Cisco hardware & believe that they Can be more cost efficient than AWS/Azure.'

In my experience, it's a combination of fear, bad math, effective sales reps and a strong sense of self-preservation.

There are undoubtedly places that self-hosting is more efficient or a simply superior option, but the typical enterprise is unlikely to fit that bill.

Trying to enact change in enterprise IT is like playing the Game of Thrones. There are all sorts of entrenched interests and alliances at play.

Even as CTO, shifting a decent sized environment to the cloud surely costs your reseller more in margins, and possibly your account manager more in commissions than you make in a year. It also likely means cutting staff.

Regardless of their level, people get very nasty when you 'mess with their money.

It's a hell of a lot easier to just keep up the status quo and keep everyone happy. Furthermore, when the shift becomes truly unavoidable, the people powerful enough to have made the move earlier aren't going to be the ones losing their jobs.

What's really tragic is that recent history suggests that the shops which are slowest to transition will be eaten by their current allies in the end.

All of the big enterprise companies will gladly sell the executives on managed services - conveniently funded by removing entire swathes of staff.

Given a significant number of companies are most cost effective than AWS/Azure/Google/Etc with relatively stable workloads, I think you are completely nuts.

AWS's calculator and the assumptions the threw into it pretty much proves they aren't cost effective except compared to other cloud providers. They basically 'made up' the self hosted numbers. At $DAY_JOB, our self-hosting costs at a colo facility are something like 33% the cost of using AWS.

Perhaps at current AWS prices your self-hosting is 3x cheaper.

But thinking about fundamentals, a few years out, why do you think self-hosting will still be cheaper? What is it in self-hosting that economies of scale of AWS/GCE can't crush?

1) They are dependent on Intel or AMD for the higher end processors. This is a $5 billion-per-factory industry. So that is one thing they can't easily replicate.

http://www.informationweek.com/cloud/infrastructure-as-a-ser...?

They don't make that on AWS, in gross revenues. This is similarly true of most hardware inside a modern server. They can make small improvements but they can't magically take over the CPU market or any of the other core components in a 'big way'.

2) Creating all that in-house software automation is complicated, difficult, and expensive. This software isn't a requirement for in-house hosting.

So they need to crush the margins of their supplies [power; bandwidth; hardware] to the point that they can get their own margin on top of that + cover the cost of a bunch of custom in-house software development. For a stable workload...that just isn't going to happen.

AWS isn't going to cut their prices 33% after the largest price cut I've ever seen them make in April, let alone the 67% they'd need to.

Sure, in 10 years, they might manage it. I don't see it happening faster than that.

A cloud provider pays the same as you for real estate, electricity, x86 hardware, and engineering talent,

The only advantage of scale is being able to oversell, e.g. running one customers workload during their peak hours and another customer on the same kit at theirs.

Are you sure about your first sentence?

real estate - don't they explicitly look for cheap real estate and tax breaks?

electricity - cloud providers save a lot on electricity by choosing carefully their locations. Facebook built a datacenter in the Arctic, I don't see your average enterprise doing that.

- x86 hardware. They're spending a lot less by having their own custom hardware

- engineering talent. They should be saving by using local talent from elsewhere, not Silicon Valley or NYC wages. IBM does that.

http://www.slate.com/articles/technology/the_efficient_plane...

http://www.businessweek.com/articles/2013-10-03/facebooks-ne...

All companies offshore these days. Anything a cloud company can do, any large company can do.

I was discussing just this with my friend yesterday. AWS remains many multiples more expensive than managing their own hardware; additionally their main office's internet connection remains less than perfect (which is a whole other wildcard.)

The question is, what happens when the cloud is cheaper than doing it yourself?

The cloud already is cheaper. Just for some reason your belief system only recognizes AWS and ignores Digital Ocean, Linode, Rackspace, Joyent, and about a thousand other smaller companies.

Nope.

Checkout online.net vs. any of those ;)

Looks like you have an API, but not sure if there is a way to create a new server with the API?

Then people will switch, but the complexities of the software AWS, et al makes me think that it'll never beat bare metal for a relatively stable load.

It saves money by eliminating expensive infrastructure engineers. Enterprise IT spends more money on PMs than tech.

However, running on the cloud also requires cloud-specific adaptation and engineering, it's not free.

The cost comparison is usually skewed in the cloud's favor because people tend to compare "server-quality" local hardware with cheaper-than-basic-PC level hardware in the cloud. The real win here is that you don't need super fancy hardware to run fairly reliably. It's not just the hardware costs, it's that when you run "unusual" hardware, you're going to have to do more administration/training. Beyond that, the cloud's advantages are trickier.

I think commoditization of computing - which is what the cloud is all about - is a valuable trend. But the benefits of specific clouds are often vastly oversold, suggesting that it's somehow magic pixie dust that avoids all sysadmin duties; or that it's the only or best way to limit those costs. Also, there's this idea that you really, really need multiple redundant servers for ever menial website out there, because that three-hour outage of your cats pics is going to be terrible. And that presumption makes the cloud more attractive, since presumably outages there simply mean you spin up a new instance (even though that is also not always so simple).

I agree -- no magic.

One of my duties as an IT manager in a big enterprise a few years back was running an install group. 7-10 FTEs, two trucks and lots of budget for rental vehicles and expenses. Most of their job was installing servers in the field and arguing with landlords about specific requirements we had (ie, 30 amp electric service, a/c in the closet, certain types of locks, etc)

Today? The office's admin calls whomever the preferred telco partners are and gathers quotes for internet service. That group is gone.

Also, never underestimate the cost of bureaucratic waste. In the place I work in now, I can probably jot down 6-10 names whose main work functions could be replaced by pull down menus on a webpage. When you outsource parts of your business, those functions tend to go away.

In larger companies most hardware is run at about 10% capacity for about 10 hours of the day. Worse are data centers which cost 100s of millions to make and sit unused.

I've done some analysis for a large company that spends about 500 million a year on IT infrastructure. Biggest expense? People, second network, after that trails mainframe and Microsoft OS licensing as well as Oracle. Server hardware cost everyone loves to talk about but actually was quite low.

Things like AWS cut network, people and can commoditize vendors such as Oracle and MSFT. Biggest savings is on not wasting time on racking hardware.

Different orgs have different cost models, and AWS is most definitely NOT more expensive for lots of them. It's also not always about bare bottom possible price for 'colo' type hosting. There's a ton of value to bringing up capacity for a few days and then making it just go away, which many F500 and govt shops struggle with.

Hence 'relatively stable loads'.

I never said it was worthless, just for alot of cases it doesn't make sense.

> What on earth makes these enterprise companies > believe That they can take Cisco hardware & > believe that they Can be more cost efficient > than AWS/Azure.

Effective sales people and existing relationships

Reply on news.ycombinator.com