3 ms·

Server power supplies are expensive, and data centers sure do care about the bottom line. If it costs $20 to fix a $100 power supply, why would you replace it?

tinco 4y ago
How frequent are the failures? It costs a lot less than $20 to repair a broken power supply, but if you break less than $5000 worth of power supplies per month you're not breaking even on the technician not even counting the management overhead. I don't think small cost savings like that are going to be implemented simply because it makes the operation more complex without landing any manager a big fat bonus. Unless I'm grossly underestimating the failure rate of those supplies?
- nwiswell 4y ago
  I think you're underestimating the scale of the modern data center.
  - tinco 4y ago
    So you're saying in your DC enough of them break that they could feed a technician? Say the technician would have an 80% success rate and he saves $100 bucks per repair, there would need to be 63 failures per month for there to be 0 cost savings. What's the minimum cost savings for it to be worth a managers time? Maybe double it again so 125 failures per month? I wonder if the vendor would void the warranty and that would have any consequences. Does you DC get around 125 failures per month?
c_o_n_v_e_x 4y ago
Actually I did a study on this (I'm a PM for a small server company) about a year ago with data scraped from a popular US based online vendor. For PSUs >1000 watts, I found that server supplies are cheaper both on a relative (more watts / dollar) and absolute cost basis than ATX form factor supplies.
- jandrese 4y ago
  Aren't the >1k watt ATX power supplies usually the "gamer edition" ones with huge markups? They cost a lot because they are a small niche--very few PCs need power supplies that big, and the ones that do are running multiple wildly expensive GPUs so budget is clearly not a concern. There's probably more legitimate demand on the commercial side for >1kw PSUs so they can enjoy larger economies of scale.

tinco 4y ago

How frequent are the failures? It costs a lot less than $20 to repair a broken power supply, but if you break less than $5000 worth of power supplies per month you're not breaking even on the technician not even counting the management overhead. I don't think small cost savings like that are going to be implemented simply because it makes the operation more complex without landing any manager a big fat bonus. Unless I'm grossly underestimating the failure rate of those supplies?

nwiswell 4y ago
I think you're underestimating the scale of the modern data center.
- tinco 4y ago
  So you're saying in your DC enough of them break that they could feed a technician? Say the technician would have an 80% success rate and he saves $100 bucks per repair, there would need to be 63 failures per month for there to be 0 cost savings. What's the minimum cost savings for it to be worth a managers time? Maybe double it again so 125 failures per month? I wonder if the vendor would void the warranty and that would have any consequences. Does you DC get around 125 failures per month?

c_o_n_v_e_x 4y ago

Actually I did a study on this (I'm a PM for a small server company) about a year ago with data scraped from a popular US based online vendor. For PSUs >1000 watts, I found that server supplies are cheaper both on a relative (more watts / dollar) and absolute cost basis than ATX form factor supplies.

jandrese 4y ago
Aren't the >1k watt ATX power supplies usually the "gamer edition" ones with huge markups? They cost a lot because they are a small niche--very few PCs need power supplies that big, and the ones that do are running multiple wildly expensive GPUs so budget is clearly not a concern. There's probably more legitimate demand on the commercial side for >1kw PSUs so they can enjoy larger economies of scale.