Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Microsoft Wins Pentagon’s $10B Contract

nytimes.com

Post reply on HN
Print view View original

121–130 of 466 posts

Re: Microsoft Wins Pentagon’s $10B Contract

#121

Post by DuskStar »

jorvi wrote: post #102
DuskStar wrote: post #86

Earlier quoted context omitted.

Yep, and this is why electrical grid issues never take down entire regions of the US! Like the 2003 Northeast blackout, for instance. [0] Things like that, where an error in one part of the grid cascades into massive issues elsewhere - those don't ever happen. Sometimes, being big is a danger. Even when that's not the case, it can be worth accepting the same risk of each individual point on the map losing power if it…

You can be sarcastic all you want, it doesn't make you right. I never said connecting grids makes them invulnerable, just more resilient. Connecting the EU grids has made them more stable, made electricity cheaper, made energy production more efficient (no more hydro going to waste), etc. etc. There is indeed risk for cascading failure but it is very rare (how long have the US grids been interconnected vs. how many t…

You asked "Why does it make sense to have independent grids?" and I provided a reason. "But that'll never happen to us!" feels rather naive, personally - I think massive, deliberate power outages have a decent chance of occurring over the next 50 years. Future of cyberwarfare and all.

(And sometimes a system being extremely stable just means it's brittle, not strong)

Re: Microsoft Wins Pentagon’s $10B Contract

#122

Post by cromwellian »

At least it wasn't Oracle. You can't say Azure, AWS, GCP, et al, aren't highly skilled, experienced, cloud infrastructure builders. There was a lot of wining-and-dining of the Trump administration by Safra Kat which gave the impression they'd be magically awarded the contract. Thank goodness the government isn't that corrupt.

Re: Microsoft Wins Pentagon’s $10B Contract

#123

Post by DKnoll »

hn_throwaway_99 wrote: post #117
DKnoll wrote: post #109

> ...unlike, say, New York City or Detroit, whose residents found out the hard way that lots of their power comes from Canada. Interesting that this myth is still alive. It started with Bush falsely claiming the 2003 blackout was Canada's fault before an investigation had even been done. Here is the actual postmortem report. https://www.energy.gov/sites/prod/files/oeprod/DocumentsandM...

Just a request: if you intend to directly refute another's comment by posting sources, posting a 238 page document without a specific page reference isn't really helpful.

Refer to the summary of causes on page 17, as seen in the table of contents.

Re: Microsoft Wins Pentagon’s $10B Contract

#124

Post by yourbandsucks »

Scipio_Afri wrote: post #112
yourbandsucks wrote: post #81

Earlier quoted context omitted.

[flagged]

He's not anti war as much as he's anti policy that pushes up against Russian spheres of influence. You see no issue with, and in fact a lot of funding and support increase to, Saudi Arabia. Where as Obama had put a pause on things during the initial outbreak. Didn't condemn things outright in public, but the Obama administration was definitely taking a step back during the Yemen war. The trump administration also has…

I'm a 2x Obama voter, but Obama totally greenlit and supported Saudi Arabia in Yemen from the beginning. Things were getting hot there in like 2010, this spanned almost his entire tenure. That's another area where as soon as things change hands, everyone has a different opinion depending on if it's the red guys or the blue guys.

As far as hooking Russia/Turkey up, Syria would be a better example IMO. We ceded Syria, the Saudis are still 100% ours. In either case, part of being non-interventionist means ceding things to regional powers.

Re: Microsoft Wins Pentagon’s $10B Contract

#125

Post by Scipio_Afri »

DuskStar wrote: post #86
jorvi wrote: post #75

Earlier quoted context omitted.

Why does it make sense to have independent grids? In Europe most of our grids are already connected and there are plans to create a supergrid that stretches all the way from Finland down to North Africa (mainly to get connected to the future Saharan solar plants). Usually the bigger the grid the more resilient it is and the less dispersion there is in demand.

Yep, and this is why electrical grid issues never take down entire regions of the US! Like the 2003 Northeast blackout, for instance. [0] Things like that, where an error in one part of the grid cascades into massive issues elsewhere - those don't ever happen. Sometimes, being big is a danger. Even when that's not the case, it can be worth accepting the same risk of each individual point on the map losing power if it…

Well the Northeast Blackout of 2003 had quite a few lessons learned, but the fact of the matter is over time the US grid has become more interconnected - just like Europe - and through the interconnections brings more stable and cheaper power. Sure one off events happen, but the grid is more stable now than in the past when it was less isolated. Part of that is high voltage DC to DC links that allow for sharing of electrical over production, while at the same time isolating grids. Even within ISO grid regions these links are used.

The main benefit from having one ISO in the US would be a reduction of rules. If you wanted to build solar or gas or wind in California, it would be subject to the same grid rules as Texas (e.g. how to handle a fault on a line - how quickly it recovers from fault, from frequency or voltage deviations, etc.). The fact that a particular power grid is isolated from others means they can't benefit from cheaper power due to overproduction of say wind power in Kansas. That electricity goes to someone else or goes to waste.

Finally if anyone here cared to read, the FERC doesn't regulate the Texas grid.. it's ISO has its own rules (which I'm pretty sure are similar to the rest of the ISOs in the US which are subject to meeting the FERC regulations or going beyond them). The Texas grid is connected to other grids.

"The Texas Interconnection is tied to the Eastern Interconnection with two DC ties, and has a DC tie and a VFT to non-NERC systems in Mexico. There is one AC tie switch in Dayton, Texas that has been used only one time in its history (after Hurricane Ike).

On October 13, 2009, the Tres Amigas SuperStation was announced to connect the Eastern, Western and Texas Interconnections via three 5 GW superconductor links.[2] As of 2017, the project was reduced in scope and only related infrastructure was constructed for nearby wind projects connecting to the Western Interconnection." https://en.wikipedia.org/wiki/Texas_Interconnection

Re: Microsoft Wins Pentagon’s $10B Contract

#126

Post by rgrieselhuber »

delecti wrote: post #98
rgrieselhuber wrote: post #71

Earlier quoted context omitted.

Why is that weird?

It was kinda a whole todo the last time some people tried to secede from the Union. It's weird to think it's a good idea to try again.

Texas is a special case in this context, although highly contentious.

Re: Microsoft Wins Pentagon’s $10B Contract

#127

Post by Cyph0n »

p0rkbelly wrote: post #116
slenk wrote: post #106

Earlier quoted context omitted.

But the integration between the two helps...

What does my word processor have to do with my GPU cluster I use for creating ML models?

Did you not install the Word add-in that offloads building the table of references to a GPU cluster?

Re: Microsoft Wins Pentagon’s $10B Contract

#128

Post by mxcrossb »

jeffwass wrote: post #3

The key point in the briefing (and the “developing” article is fairly brief) is that Amazon was considered the front-runner for contract, but Trump has a personal vendetta against Amazon/Bezos and also made comments that he’d get personally involved in the JEDI decision (Joint Enterprise Defense Infrastructure). So of course this invites the question of whether petty politics influenced a $10B govt deal.

Was this info not in the original article? It seems to be the most important point, yet nobody else is discussing it in the comments.

Re: Microsoft Wins Pentagon’s $10B Contract

#129

Post by metaphor »

z3t4 wrote: post #48

What do the military need 10 billion US dollars in cloud computing for ? Are they going to fork Bitcoin? DDOS attack? If they are for example going to archive video footage, it's probably much cheaper to print it to film.

https://news.ycombinator.com/item?id=20427578

Re: Microsoft Wins Pentagon’s $10B Contract

#130

Post by Scipio_Afri »

DuskStar wrote: post #121
jorvi wrote: post #102

Earlier quoted context omitted.

You can be sarcastic all you want, it doesn't make you right. I never said connecting grids makes them invulnerable, just more resilient. Connecting the EU grids has made them more stable, made electricity cheaper, made energy production more efficient (no more hydro going to waste), etc. etc. There is indeed risk for cascading failure but it is very rare (how long have the US grids been interconnected vs. how many t…

You asked "Why does it make sense to have independent grids?" and I provided a reason. "But that'll never happen to us!" feels rather naive, personally - I think massive, deliberate power outages have a decent chance of occurring over the next 50 years. Future of cyberwarfare and all. (And sometimes a system being extremely stable just means it's brittle , not strong)

The word you're looking for is resilience. That is basically all the regulators and grid ISOs do for the most part, make sure the grid is stable with different power sources of all types - be it gas generators and wind turbines that exist within a region under their control, or connections to outside grids. Regulators, national security people and others are definitely looking at what a hacking event might look like. The various grids, connected to each other or not, are definitely not as well designed to be resilient to something like a deliberate malicious act to damage power equipment by essentially hacking into the control systems to do something that they never were intended to do.

The design of electrical power grids is such that their control systems are by their very nature suppose to keep frequency and voltage within tolerances no matter the unpredictability of the grid. This is the definition of stability and resiliency that you're talking about. And the design takes into consideration failures of any device on the grid.

That being said, even being a minor power grid, it isn't actually isolated. Its just not subject to FERC regulations. But for the most part, its own regulations are similar to that of what other grid operators are doing, because it makes sense to do for grid stability reasons.

"The Texas Interconnection is tied to the Eastern Interconnection with two DC ties, and has a DC tie and a VFT to non-NERC systems in Mexico. There is one AC tie switch in Dayton, Texas that has been used only one time in its history (after Hurricane Ike).

On October 13, 2009, the Tres Amigas SuperStation was announced to connect the Eastern, Western and Texas Interconnections via three 5 GW superconductor links.[2] As of 2017, the project was reduced in scope and only related infrastructure was constructed for nearby wind projects connecting to the Western Interconnection."

Post reply on HN