Is a billion dollars worth of server lying on the ground? • Hacker News Board

Re: Is a billion dollars worth of server lying on the ground?

#81

Post by bogwog » Mon, Nov 02, 2020, 2:39 PM UTC

iampims wrote: post #71

I’m surprised no one has mentioned the number of data centers that AWS offers. Good luck using OVH anywhere but France/US east coast.

If all you need is a few VMs, then there's no real reason you have to be stuck on a single provider. If OVH only has servers on the east coast, then you find another provider for the west coast. I've never used OVH, but I doubt creating an account with them (and other providers) is difficult. Shopping around is the only way to get the best prices.

Maybe it's mildly annoying to have to deal with multiple hosting accounts, but that's a small price to pay for potentially huge savings, IMO.

Re: Is a billion dollars worth of server lying on the ground?

#82

Post by hbogert » Mon, Nov 02, 2020, 2:40 PM UTC

iampims wrote: post #71

I’m surprised no one has mentioned the number of data centers that AWS offers. Good luck using OVH anywhere but France/US east coast.

Why? They have 8 PoPs stretching from US west coast to Singapore and Australia. That seems pretty okay.

Re: Is a billion dollars worth of server lying on the ground?

#83

Post by WrtCdEvrydy » Mon, Nov 02, 2020, 2:41 PM UTC

mattmanser wrote: post #64

WrtCdEvrydy wrote: post #52

Earlier quoted context omitted.

AWS allows you to shift your costs from CapEx to OpEx. Companies with low CapEx are valued higher since "theoretically" you could remove that bill by moving to another provider. Financial Engineering is just another part of software engineering and the cloud enables it.

This is not a real saving in my experience. The DevOps time for an app is so trivial. I actually just setup a .Net core app on Linux/mysql. My Linux experience is old and very limited. I have used AWS for years for other things (S3, cloudfront, transcribe, etc.). Initially I setup an elastic beanstalk app/separate mysql instance on my own AWS account just so I could quickly deploy (all new to me). Then I setup the ap…

To be fully honest, for personal work, I use Caprover for DevOps.

Edit: The move from CapEx to OpEx is not about savings, it's often about shifting the costs in your books.

Re: Is a billion dollars worth of server lying on the ground?

#84

Post by manquer » Mon, Nov 02, 2020, 2:41 PM UTC

sgt wrote: post #63

kuon wrote: post #13

With the price of pro fiber (redundant with SLA) I recently moved some apps back to our own servers in house. This did cut the price down dramatically. I would not recommend it for super critical apps (except if you have your own state of the art data center, but I am not speaking about that), but having 5-10 servers in a secure cabinet will give you infinite flexibility for little cost. We pay around 5000$/y to keep…

I like the idea, but remember that the SLA with your fiber provider doesn't mean your fiber magically will come back within a few hours when someone in the street cuts through the cable by accident. Your clients will still experience downtime unless you have two fiber lines going in different directions - completely redundant. If you don't have that, I would recommend some kind of a licensed microwave link on your ro…

You always have a DR offsite, even if you are hosting on Big Cloud that would be sensible to setup.

Your RTO and RPO needs will dictate your DR setup in both scenarios. Most apps can take few hours hit if the alternative is spending 2-3x .

Re: Is a billion dollars worth of server lying on the ground?

#85

Post by ti_ranger » Mon, Nov 02, 2020, 2:43 PM UTC

jwr wrote: post #21

When determining what to use for development of my SaaS, I did a comparison of what you actually get from providers. The full article is at https://jan.rychter.com/enblog/cloud-server-cpu-performance-... My takeaways were that many cloud provider offerings make no sense whatsoever, and that Xeon processors are mostly great if you are a cloud provider and want to offer overbooked "vCPUs". I haven't tested those specif…

> When determining what to use for development of my SaaS, I did a comparison of what you actually get from providers. The full article is at https://jan.rychter.com/enblog/cloud-server-cpu-performance-...

Your results (e.g. that z1d.xlarge with 4 vCPUs is only 10% slower than z1d.2xlarge with 8 vCPUs) shows that the "performance" you were testing was disk IO throughput (probably dominated by disk latency), not vCPUs.

> My takeaways were that many cloud provider offerings make no sense whatsoever, and that Xeon processors are mostly great if you are a cloud provider and want to offer overbooked "vCPUs".

> I haven't tested those specific setups, but I strongly suspect a dedicated server from OVH is much faster than a 4.16xlarge from AWS.

You seem to be implying that AWS/EC2 does CPU over-provisioning on all instance types; this is incorrect, only T-family instance types use CPU over-provisioning.

Re: Is a billion dollars worth of server lying on the ground?

#86

Post by twic » Mon, Nov 02, 2020, 2:45 PM UTC

cthalupa wrote: post #67

twic wrote: post #49

Earlier quoted context omitted.

> It's easier to just continuously benchmark performance and kill machines that don't perform as expected. It's very easy to automate. It's also work you simply don't have to do if you rent hardware.

You don't think there is variance in performance when you rent hardware? You've never had an HDD or SSD that underperformed vs. others of the same type or model? You've never had a stick of RAM that underperformed, or threw lots of CEs? I'll be frank: If you haven't seen performance variance with physical hardware, you either have been incredibly lucky, not paying attention, or not working at a very large scale.

Ah, that delightful cloud propaganda move, where you assert that because some undesirable thing happens to both VMs and physical machines, that they're the same, elegantly glossing over the orders of magnitude difference in severity.

Re: Is a billion dollars worth of server lying on the ground?

#87

Post by ti_ranger » Mon, Nov 02, 2020, 2:52 PM UTC

bsenftner wrote: post #47

Cthulhu_ wrote: post #40

Earlier quoted context omitted.

From AWS' own marketing propaganda: How much would said clients have to invest in people and hardware otherwise? What if their application becomes an overnight success and needs to scale up fast? Sure, if it's an established company then using their own hardware (and people to set up and manage it) might make sense; iirc Dropbox is a fairly recent big player that made that move. But otherwise it's a big upfront inves…

For my own startup, I built a small cluster of 17 servers for just beneath $55K, and that had a month-to-month expense of $600 placed in a co-lo. In comparison, the same setup at AWS would be $96K per month. And it is not hard, easy in many ways. Do not be fooled, the cloud companies are peddling is an expensive scam.

> For my own startup, I built a small cluster of 17 servers for just beneath $55K, and that had a month-to-month expense of $600 placed in a co-lo. In comparison, the same setup at AWS would be $96K per month.

Why would you build exactly the same setup in AWS as for on-prem, unless your objective is to (dishonestly) show that on-prem is cheaper?

Lift-and-shift-to-the-cloud is known to be more expensive, because you aren't taking advantage of the features available to you which would allow you to reduce your costs.

Re: Is a billion dollars worth of server lying on the ground?

#88

Post by cosmie » Mon, Nov 02, 2020, 2:54 PM UTC

ronyfadel wrote: post #72

cosmie wrote: post #70

Earlier quoted context omitted.

"in house" doesn't necessarily mean in a residential home; it refers to "on premise" more generally. A previous company of mine did the same thing - they converted a maintenance closet into a server closet. Even with renovation costs to improve ventilation and electrical load to support the use case, it worked out substantially cheaper than cloud hosting. A few things we ran on it: - A large data infrastructure. We h…

Thank you for the info. I got carried away mistaking “in house” for running a business (and hardware) from home!

To add, although not quite the same as running a full server closet, I've also repurposed an old laptop as an always-on server and run a handful of web services for both professional and personal purposes from my home connection:

- Some utility applications I maintain for consulting clients. These tend to be incredibly low volume, accessed a handful of times a month at most.

- I host some analytics infrastructure for clients. It's for reporting, rather than data capture, so latency and uptime aren't super critical.

- I run a personal zerotier[1] network, which I use as both a virtual LAN across all my devices hosted everywhere as well as to tunnel my laptop and mobile traffic when I'm not at home. My internet gateway is hosted at home, so all my mobile and public wifi traffic routes through my home connection.

- I do a minor bit of web scraping. This is fairly low volume, and a mix of both personal and professional work. If it was higher volume I wouldn't host it at home, due purely to the risk/potential hassle of IP blocks and complaints.

I have a fairly stable symmetrical 1Gbps fiber connection (plus a block of static IPs). It's a residential connection so has no firm SLA, but still achieves ~500Mbps when experiencing "congestion" (in whatever form that means for fiber to the premise) and has only been down twice (once during a total power outage and another an attempt to use the parental control features for one device resulted in every device being MitM'd by an invalid SSL cert).

I also have a mobile hotspot I use when traveling for work, which I have connected as a failover ISP when I'm at home. This covered both instances of downtime I've experienced. And in the case when it doesn't and a client needs to access a service, I maintain external backups that I can spin up on VPS somewhere in a pinch. Probably not enough guarantees for a primarily SaaS based business, but has worked without a hitch when user-facing SaaS websites/apps are only a minor secondary component of the work.

[1] https://www.zerotier.com/

Re: Is a billion dollars worth of server lying on the ground?

#89

Post by ls65536 » Mon, Nov 02, 2020, 3:04 PM UTC

aliceryhl wrote: post #4

I run my personal website and other small projects on VPSs I bought from OVH, and they're really cheap and work very well. Can recommend.

I agree. They're really great if you need well-priced, no-hassle infrastructure. As others have already mentioned, they may not have some of the higher-level features that AWS comes with, but if you don't need these, then there's no reason to be paying more for it.

I'm really fascinated by their approach to building datacenters, which seems to include taking over disused industrial sites and converting them for datacenter use. For example, their datacenter just outside of Montreal is on the site of an old aluminum smelter [0]. In this case, I'm sure the proximity to plentiful and cheap hydroelectric power nearby (as would have been beneficial for aluminum smelting in the past) was a major factor in the choice of location as well.

[0] https://baxtel.com/data-center/ovh-beauharnois-quebec-bhs

Re: Is a billion dollars worth of server lying on the ground?

#90

Post by mwcampbell » Mon, Nov 02, 2020, 3:06 PM UTC

Spivak wrote: post #76

KingOfCoders wrote: post #18

Earlier quoted context omitted.

Many people use AWS because everyone uses AWS. Many of my clients have no need for AWS but still use it, at least until the VC money runs out. Usually then I have to go first, then servers are moving to somewhere cheaper when a new CFO comes in.

At my company we colo at a few local datacenters and have do deal with a huge amount of pressure from our investors as to why we're not using AWS. The points that always seem to come up: * AWS is a known quantity and it's easier to evaluate our business with it. * AWS provides "outage damage control" because AWS outages make the news and customers are more understanding. When our ISP has issues it just looks bad on u…

How do you do failover if a server fails or if connectivity to one of those datacenters is lost? With AWS I could just set up a multi-availability-zone RDS deployment for the database and an auto-scaling group for the web tier and be confident that AWS will recover the system from most failures. To me, that is the major selling point of any of the hyperscale cloud providers.