5 ms·

disagree. There comes a certain scale where doing it yourself is way cheaper. There's other reasons to outsource to Amazon than just cost though

all_blue_chucks 9y ago
There is way more that goes in to operating a large scale computing infrastructure efficiently than the cost of hardware. Like way, way more. If anyone could replicate it at the right scale then every large organization would be competing with Amazon for the cloud infrastructure market.
- toomuchtodo 9y ago
  Is this your opinion because you work at Amazon? Or because you can prove this to be the case?
  - all_blue_chucks 9y ago
    Can you prove your case? If you are correct it's just a matter of scale, it stands to reason that any large organization would be muscling in on the high-margin cloud computing business. That clearly is not the case. But I'm happy to see any evidence you have to the contrary.
- halbritt 9y ago
  Way more? Nah. I ran a couple data centers on either coast each burning about half a megawatt, which isn't that much in the grand scheme of thing. My employer at the time offered a SaaS product to Fortune 100 companies. The product required a realtime kernel and QOS in the network, neither of which is supported in AWS. I did a fair bit of math comparing costs. Typically, you could pay off the CapEx with 6 months of usage in AWS. IIRC we had around $8m in salaries for the ops org. I'm not sure we would've needed a great deal fewer people having all that infra in AWS. The real problem comes from trying to find a team of people willing to engineer your service for failure. That kind of attitude just doesn't exist in the world of IT, so anyone would be hard pressed to put a team together. Good luck finding a network engineering team that knows BGP well enough for such a thing. Good luck finding a team of engineers that know Openstack well enough to keep it running. Good luck finding folks to run Ceph. Good luck getting all that right. If you could, though, there are certainly savings to be had.
  - all_blue_chucks 9y ago
    We aren't talking about a datacenter to support one product. We are talking about an entire platform for building and operating any number of products on top. It isn't comparable.
    - halbritt 9y ago
      I referred to the two data centers I had in the US. How many data centers do you think Oracle operates in the US, presently?
Rebelgecko 9y ago
There's a certain scale where it is definitely possible to do it cheaper, but the government isn't really able to do that. Even if the program didn't crash and burn (fairly likely- there's plenty of examples of billion dollar government software programs that turned into vaporware), it'd be taking on a ton of risk just to set up the same thing without the AWS profit margin (maybe 10-20%?)
coredog64 9y ago
Sometimes there are diseconomies of scale as well. Like all of a sudden you have to put a data center into an otherwise shitty location because the committee chairman needs to show some results to constituents. Or two departments have essentially the same requirements, but they compete for budget and each have to justify their staffing by duplicating effort. (These pathologies aren't unique to government. You can also find them in large private companies.)
- cobookman 9y ago
  "There's other reasons to outsource to Amazon than just cost though" I think that falls under the `other reasons`

all_blue_chucks 9y ago

There is way more that goes in to operating a large scale computing infrastructure efficiently than the cost of hardware. Like way, way more. If anyone could replicate it at the right scale then every large organization would be competing with Amazon for the cloud infrastructure market.

toomuchtodo 9y ago
Is this your opinion because you work at Amazon? Or because you can prove this to be the case?
- all_blue_chucks 9y ago
  Can you prove your case? If you are correct it's just a matter of scale, it stands to reason that any large organization would be muscling in on the high-margin cloud computing business. That clearly is not the case. But I'm happy to see any evidence you have to the contrary.
halbritt 9y ago
Way more? Nah. I ran a couple data centers on either coast each burning about half a megawatt, which isn't that much in the grand scheme of thing. My employer at the time offered a SaaS product to Fortune 100 companies. The product required a realtime kernel and QOS in the network, neither of which is supported in AWS. I did a fair bit of math comparing costs. Typically, you could pay off the CapEx with 6 months of usage in AWS. IIRC we had around $8m in salaries for the ops org. I'm not sure we would've needed a great deal fewer people having all that infra in AWS. The real problem comes from trying to find a team of people willing to engineer your service for failure. That kind of attitude just doesn't exist in the world of IT, so anyone would be hard pressed to put a team together. Good luck finding a network engineering team that knows BGP well enough for such a thing. Good luck finding a team of engineers that know Openstack well enough to keep it running. Good luck finding folks to run Ceph. Good luck getting all that right. If you could, though, there are certainly savings to be had.
- all_blue_chucks 9y ago
  We aren't talking about a datacenter to support one product. We are talking about an entire platform for building and operating any number of products on top. It isn't comparable.
  - halbritt 9y ago
    I referred to the two data centers I had in the US. How many data centers do you think Oracle operates in the US, presently?

Rebelgecko 9y ago

There's a certain scale where it is definitely possible to do it cheaper, but the government isn't really able to do that. Even if the program didn't crash and burn (fairly likely- there's plenty of examples of billion dollar government software programs that turned into vaporware), it'd be taking on a ton of risk just to set up the same thing without the AWS profit margin (maybe 10-20%?)

coredog64 9y ago

Sometimes there are diseconomies of scale as well. Like all of a sudden you have to put a data center into an otherwise shitty location because the committee chairman needs to show some results to constituents. Or two departments have essentially the same requirements, but they compete for budget and each have to justify their staffing by duplicating effort. (These pathologies aren't unique to government. You can also find them in large private companies.)

cobookman 9y ago
"There's other reasons to outsource to Amazon than just cost though" I think that falls under the `other reasons`