4 ms·
Secretly Throttling Is Never Okay
- samstave 5y agoI have had so many experiences with this problem across many carriers ASIDE:: (A tiny flex: I was the IT DIR at DECIDE.com in ~1999-/+ and we built a system to measure cell signal quality such that a consumer would be able to make a decision on a carrier's ability to maintain signal btwn P.A and P.B... Interestingly, there was a a graphic designer named Brian Thorp - and he came up with the first 'magnifying' glass as an icon for search that I was aware of - I could be wrong, but he made that in ~1997 and it was the first icon I had seen of such... (Ill go search my memory) --- ANYWAY - the point is that I have seen secret throttling happen across MANY (ALL) carriers in the following theft strategy: "OFFICE SPACE" They arent slurping your 'data package' in real-time-transfer... its the Chinese trickle method... When the Chinese air-gapped Lockheed via a few suppliers in Taiwan, they were trickle hacking. They didnt try to firehose. But when Lockheed found out about this, there were several security measures that went into action. Most note: TLS across the board. and Delete ANY email older than X. . (they already had a mitigated exit surface with only 3 egress... But the Chinese were aggressive on their phish-nets. USBs Dropped about... and more importantly, plants in def conferences... whereby agents were phished with an email a few weeks after the conf "Hey we met at conf X email me here at Y" and exploits abound... trickle trickle... until you get caught - then do the firehose. until cut. Cell carriers are *trickle charging you for data when you are attached to wifi so that you dont recognize that you are being robbed.* --- YOU DATA IS BEING TRICKLE CONSUMED EVEN WHEN YOU ARE ON WIFI' 'Setup a device connected to your router. Setup a burner phone to connect to the segregated router. Browse on WIFI whilst you still have cell connected.... Browse the exact same from a diff device that is wifi only. Compare usage.... --- The other thing I find not talked about AT ALL, would be the HDDs from [Sandia?,JPL?, Livermore?, etc?] -- The fn HDDs that were later recovered and supposedly found behind the copiers/MFPs, and supposedly there were some arrests... but no follow up. --- Then there was the Chinese spy couple that the husband worked for [JPL?] and they lived this really weird minimalist existence (mattress on the floor, horror movie lighting fixtures, nothing in the fridge) -- and he was a nuke designer for [JPL?]... --- My grandfather was a Nuke Designer for GE his entire career and helped design and build Hanford. (I have a lot of nukes in my family) - Every single one of them were the most organized, meticulous engineers I have ever known... Their notes, their books,... No FN way this JPL chinese spy didnt have EXTENSIVE JPL info he was feeding.
- therealcamino 5y agoThe article is about CPU scheduling.
- gumby 5y agoI think people in general value battery life over raw performance and I think the manufacturers have interview and survey data to prove it.
- Anunayj 5y agoyes, I somewhat agree, I've seen my laptop often clock up using minimal use, where I'd just like it to run on it's lowest frequency. (This usually happens when the laptop's on power, charging). A quick power plan change fixed that, but it'd have been nice if it was that way.
- smabie 5y agoThis article seems to be assuming that it's not okay and then concluding that it's.. not okay. I personally don't find it very convincing. That being said, these manufacturers should probably have a max performance toggle so that users have some choice in the matter. I think it's still fine to throttle by default.
- NobodyNada 5y agoI can see where the author is coming from — the line between throttling and planned obsolescence seems thin (especially when devices with worn-out batteries are subject to increased throttling). However, throttling serves a very important purpose and is much more prevalent than the author seems to realize. Every processor will throttle itself when it gets hot (and I can even remember Apple talking about this in a keynote where they were comparing different models of iPhone by “minutes before throttling at 100% sustained CPU load”). I don’t think users need a setting would allow their device to melt rather than throttle. Similarly, the “Batterygate” throttling was arguably in the best interests of the customer because a phone that performs poorly is more useful than a phone that doesn’t work at all. I don’t really care if OnePlus or whoever decides to slow down Facebook a little bit, as long as they don’t pull a Volkswagen and send un-throttled hardware to reviewers or anything like that. The engineers decided the phone was faster than necessary but was lacking in battery life, so they decided to make a tradeoff just as if they had decided to switch out the physical hardware for a more efficient processor. The reviews will say “Facebook is a little sluggish but battery life is great”, and whether that’s because of hardware capabilities or throttling is one of many implementation details that I don’t need to be concerned about. While a “performance” vs “efficiently” toggle would be cool and potentially useful for, I would consider that an extra feature and not a hard requirement. In fact, Apple has this feature already with the “low-power-mode” toggle that applies additional throttling and reduces background activity and network traffic.
- Dylan16807 5y ago> Similarly, the “Batterygate” throttling was arguably in the best interests of the customer because a phone that performs poorly is more useful than a phone that doesn’t work at all. > as long as they don’t pull a Volkswagen and send un-throttled hardware to reviewers or anything like that. Well the problem there is that you couldn't know that the experience shown in reviews would disappear over time. It's not expected for working hardware to perform worse and worse over time. > The reviews will say “Facebook is a little sluggish but battery life is great”, and whether that’s because of hardware capabilities or throttling is one of many implementation details that I don’t need to be concerned about. Reviewers only have so much time. It shouldn't be "buyer beware, sorry if the reviewers didn't find every land mind that affects you". > Volkswagen Did they even send different hardware? The trigger conditions for the cheat were so precise that they didn't need to. Surely that wouldn't excuse the cheating though, right?
- KMnO4 5y ago> In recent years, the performance of smartphone SoCs has reached a point where their power is often overkill in certain scenarios I wish this didn’t happen. If everyone had a slower phone CPU: 1. Apps would have to be more efficient 2. That means people with slow phones would get a better experience 3. People with faster phones would also get a better experience It’s crazy that you can run C programs that have been compiled into JS and run on an interpreter written in a language that runs on a JVM, which consumes 10,000x CPU cycles than native and the solution is “more cores, more RAM, bigger battery “.
- SahAssar 5y ago> C programs that have been compiled into JS and run on an interpreter written in a language that runs on a JVM If you are talking about wasm, then i'd point out that wasm isn't js, the runtime generally isn't a pure interpreter and it usually does not run on a jvm.