The EeePc's 36W full load is quite similar to the C7's 55W, recalling that the latter has two 3.5" drives and various fans rather than SSD. If a casual browsing machine were what's required, they'd both be fit for purpose following de-moving-part-ification of the C7.
Others have mentioned laptops (including the Mac Mini, which is just half a laptop) as requiring less power. Well, yes - the aim isn't to find the lowest power machine with storage and net connectivity, as then my 4 AA cell powered ARM-based HP 50G calculator would shine. For it takes SD cards and has a TCP/IP stack written for it:-). Actually, the 8088-compatible Psion Series 3a would do better: it has a nicer keyboard than most palmtops and a screen which has lasted two decades without deterioration. Two AAs, *two* flash slots - albeit proprietary.
There is absolutely no way that you're reading 5W at the wall with one Core 2 Duo plus two spinning drives, sorry. If one of these is a 7200rpm+ desktop drive, 5W+ will be needed to spin one drive. I'm running 4 year old but stably performing WD SE16 drives, and I know I could do better, but not this well.
Which motherboard, drives, PSU? Is everything spinning?
I'm not surprised to hear that idle performance is excellent on a modern desktop CPU, and I've heard similar achievements with a single desktop HDD. The 45mm Core 2 Duos also idle very efficiently, but shoot up under load with no hardware acceleration to mitigate.
I know that I'm unlikely to get readings much below current unless I switch to a PicoPSU or similar, because inefficiencies in a desktop PSU shoot right up at very low power output.
Wotcha. I'm glad the thread's received quite a bit of interest!
(1) Why "should" disks be spun down on an idle server? It's not like CPU which can pretty much instantly switch between power modes: an HDD takes time to spin up and spin down, and its life is decreased in doing so. You may decide to spin down drives, but only after a sufficient period of idling. If you really want to save power, an idle server "should" be asleep with WakeOnLan enabled.
(2) The BeagleBoard has 128MB of what looks to be on-chip memory, which is of course going to be lower power than if you need to route to a separate connector and fit off-the-shelf DIMMs. But it is an unreasonably small amount for a server, which benefits from a large disk cache - especially when it needs to decrypt what it reads from the drives but can keep memory unencrypted.
(3) The 256MB NAND flash might be useful for a root filesystem (what's the wear-levelling algorithm on this thing?), but it's hopeless for server storage. It's true that flash is going to idle much more efficiently than an HDD, and it's one of the things I raised in OP.
(4) Could you give an example where the "Atom CPU when idle" is consuming more power than a BeagleBoard "system"? I'm not sure what sort of system you mean, but I'm obviously talking about a NAS.
That 50W is three orders of magnitude greater than 50mW is entirely irrelevant. The draw of the CPU/mobo isn't so dominant in a typical system, and you're comparing one comparatively low horsepower CPU on a low feature board with a more powerful CPU on a board with enough expandability to be applied as a small desktop/server.
The BeagleBoard is still cool, of course, but for other things. And if you compared power draw of a bare VIA mobo with the BeagleBoard, I'd expect the latter to win out.
For CPU cooling, I have replaced the stock PC2500e fan with a 40mm Rasurbo specced at 0.6W max, which I've then further slowed down. One of the dedicated case fans is used to cool the drives (quite effectively, so it'd be there regardless of CPU), and the other may be extraneous, as the PSU draws out air. It is left over from when I had a higher power board in there, particularly because I was considering a Zalman ZM-NB47J passive cooler for the C7 CPU and then keeping that case fan to ensure good airflow over it. This'd be quieter.
Many people seem to run completely motionless Via C7 boxes, including by modding this motherboard - although more likely is that you'd start off buying a fanless board. They're commonly available clocked at 1GHz, though 1.2-1.5GHz fanless seems possible.
All this said, I'd be interested in recommendations for ARM mobo/CPU combinations suitable for homebrew NAS/gateway applications, and to hear about crypto performance on these boards.
Are hp still emulating the saturn processor on their arm devices?
Yes, although some operations have been accelerated with a native rewrite. But the emulation allows a very mature product to be executed without the cost and bugs inevitable with a complete port. People don't use calculators for their horsepower, but for usability and portability. The HP RPN + CAS are very usable.
FWIW, I very quickly gave up learning Saturn assembler when I still had a real Saturn device:-). Today, there are well-known ways of escaping to native ARM.
Erm, no, for developers there have been Linux kernel crypto modules supporting the Via Padlock included since 2.6.11, and if you don't want them, you can always use the crypto instructions directly. The Java is just an API option provided by Via.
45-50W is including UPS, PSU and everything spinning (and miscellaneous peripherals omitted from the above list, e.g. second Ethernet PCI, floppy, DVD drive, external modem), and with a discussion on how this is far from the lowest power solution and how I could improve things. The drives spin using around 8W each, for example.
The BeagleBoard draws 50mW at 5V DC with nothing connected, and would be hopeless for the NAS (and many other) applications described because it lacks the various ports from SATA to Gb NIC (or PCI for such expansion) which a higher power chipset provides, and lacks the horsepower necessary to make up for the lack of hardware crypto acceleration. USB NIC, now?
I had an ARM-powered primary desktop between 1990 and ~2000 (Acorn A3000, RISC PC), and now have an HP 50G calculator running off 4 AA batteries with a faster ARM CPU. I do appreciate the architecture and power efficiency of ARM designs.
Sorry - although I mentioned that the Via's great for a home/office encrypted NAS, I perhaps wasn't clear that this is precisely the application I was talking about. I was just expanding the discussion on power-efficient CPU applications, and implying that, when considering energy efficiency, a low power CPU with dedicated circuitry for popular complex operations might be the way forward.
Yes. The bottleneck for ssh probably becomes the HMAC - this thread is enlightening. SHA HMAC is afaict considered difficult to accelerate with Via's implementation, though a patch may exist. With a stock (AES-accelerated) build, software MD5 is quicker.
Wait, what? I was talking about a specific application in the general topic of power-efficient CPUs. I've yet to see a comparison of current high end Intel vs Via, e.g. Atom D510 vs Via Nano. If you genuinely need to sort all day, you probably have the intelligence and resources to prepare yourself, so you might be better off building a hardware sorting algorithm in an FPGA. You're unlikely to cluster low-power CPUs on off-the-shelf motherboards because your interconnect will be shit.
Also, what is the 1GHz referring to? Modern Via chips typically run up to 2GHz, and are architected differently from Atoms, so a straight MHz comparison is (as always) meaningless. For example, the Atom is (afaict) still an in-order CPU.
The Via C7/Nano seems to be a great chip for a home/small office server, what with its built-in AES encryption making it faster than even a high end Xeon without hardware acceleration. My current setup consists of 2*WD SE16 hard drives, APC UPS, 80+ Corsair PSU, PC2500e Nano mobo with 1GB, and a couple of 80mm case fans, together running under 50W idle, and only 7W more at full CPU load. If I were to replace the Corsair with a fanless PSU good up to 80-120W I might get an extra 5-10% efficiency; I could wipe out the case fans probably with no problem (2-3W, say), especially if I replaced hard with spinning solid state storage, and that of course would shave off around 15W. Substitute a large fanless heatsink for another W (or just get a fanless motherboard/CPU in the first place). But even as-is, it's a good improvement on my previous regular desktop CPU-based setup.
For something which is on 24 hours a day, going several months between reboots and stressed only in the IO and encryption departments, I see no reason to use a full-power desktop processor. So, what problems have you guys encountered which has meant you haven't ended up with this option?
Honestly, I've never understood why people, when faced with an idea they don't like or can't understand, just fall back to deriding the source of the idea.
I dislike the idea, *and* I don't trust the source of the idea. These two intertwined issues. Even if the idea was proposed by an entity I felt confident dealing with, I still wouldn't like the idea. Even when Google pushes good ideas, I still don't want Google implementing those ideas if they affect others on too large a scale because I don't feel confident dealing with Google. But, in this case, we have a bad idea from an entity which would be expected to propose this sort of solution (because it fits in with its modus operandi) rather than something better for the Internet as a whole.
It is perfectly legitimate to attack both the idea and its proposer's motives, just as people are free to attack both the idea/implementation of Obama's healthcare plan and question his motives. For motives will reveal any gotchas and affect the implementer's future direction. No action is performed in an altruistic vacuum, and you're regrettably naive if you think Google's final aim is "bringing IPv6 to the masses". Its final aim is to make money, and if we can't clearly see how it's doing that by pushing IPv6 as it is, it's because we lack the knowledge, not because Google are altruistic. Who knows, maybe they're negotiating with big ISPs to do something fancy with multicast? Either way, I'd like to know if they expect me to cooperate.
As a parting note, I want you to consider the difference between reasonable and Google's method in encouraging the casual enthusiast to adopt IPv6 (and every technological take-off begins with the casual enthusiast).
Expected: 1. Set up IPv6 tunnel, receive feature parity Google service via IPv6.
Experience: 1. Set up IPv6 tunnel; 2. Be surprised that you're not getting any Google IPv6 connectivity; 3. Find out that www.google.com isn't providing AAAA records; 4. Check with your tunnel provider, and find that they have signed up to some Google whitelist; 5. Assume that this means you have to make DNS requests from an IP address in their range, so get your local forwarding DNS server working on IPv6; 6. Find out that www.google.com isn't providing AAAA records; 7. Read up further, and find that the whitelist is actually a specific list of their DNS servers; 8. Find out what these DNS servers are, which involves a dig to list all the AAAA records for the local list of five or six caching DNS servers; 9. Add all the records to bind; 10. Find that you're finally getting IPv6 addresses (woohoo!), but that resolution is intermittently slow; 11. Ping each DNS server to find out which is lowest latency - they're all at least double the latency of your ISP's local DNS server, but whittle down to two; 12. After more experimentation, find that one of these servers takes around 5 seconds to respond to half the requests; 13. Whittle down to one server of reasonable speed; (14. For end-to-end IPv6, send IPv6 address for your local DNS server via RDNSS, the recent extension to the router advertisement protocol, but find that Windows 7 ignores this extension entirely. Not a great problem for me, but a barrier to those who might want to advertise the IPv6-only DNS server found in 13 directly, as they'll have to set up DHCPv6. And, if you have XP, you're SoL.)
The ironic thing here is that this describes an experience with a service not only whitelisted by Google but with administrative input from Google employees. If Google wasn't such a control freak, the user would be done at step (1).
Ah, engineers. You have the brainpower to build a Rubik's Cube solving robot, but you skip the processing which concludes that dark green on dark blue makes for an illegible web site.
(I know that sounds like a troll, but I'm interested to understand why some people with stellar abilities will nevertheless end up doing something apparently trivial like not checking that their work is legible.)
Thinking about it, since Microsoft has been onboard with IPv6 before anyone had even heard of Google, they'd be better poised to ensure a smooth transition by pushing an update which regularly checks for good IPv6 connectivity and temporarily disables attempts to resolve AAAA records if such a check fails. A simple registry/UI change should disable this service, of course.
As for where the servers are which respond to the connectivity checks - well, MS can host them and rejoice in the data it collects by hearing the IPv6 pings. They already get to phone home in their EULA for antipiracy and update checks, and the world seems to be happier with these potentially much greater intrusions.
The Linux community could produce a similar daemon. Hell, cooperate with Microsoft so you can use their ping collectors if you want, as otherwise angry rival newspaper publishers cooperate on technologies for the common good such as environmental measures. However, already having the resources to provide updates, I'm sure the Linux community can muster up the bandwidth to hear a ping every so often.
Again, the ISPs have made this bed. This guy may be one of the shining lights, but he's in an industry that's just barely competent enough to keep a regular v4 network up and alive. You can hardly blame Google for doing their diligence.
Their diligence? Google are irrelevant when it comes to pioneering either development or deployment of network layer technology. Yet instead you're telling off ISPs for not somehow not providing IPv4 properly, when they've managed to help grow the consumer Internet in the past 10 years at a rate unrivalled by probably any other consumer tech?
Google are FUDing, because it's in Google's interest to spread FUD as a precursor to the "but if you just put all the information/control in our hands, we'll make it all better..." which forms the basis of every single business decision they make.
Oh come on, quit being so dramatic. Google never rejected anyone. Did they make those guys go through a bunch of hoops? Yes, they did. But the guy flat out stated that he expects to close a peering agreement with Google.
This is because he is stubborn and persistent. Most ISPs would just stop caring before they'd even started.
It's gonna be a concerted effort between ISPs *and* content providers, and this scheme finally has a chance of getting content providers onboard.
All it does is give content providers a chance to reduce the inconvenience for under 0.8% of users once or twice (until they figure out the problem or choose the legacy URL) in the event that providers believe simply adding AAAA records is correct migration procedure. Since adding AAAA records with the current reliability of global IPv6 connectivity is a shit idea anyway - i.e. the problem is as likely to be bad IPv6 setup/connectivity on the content provider's end, or somewhere in the middle - you're trying to solve a problem by an initiative which implies a poor method.
Indeed, Google still doesn't provide usable IPv6 search for anyone outside the US, because its ipv6.google.com won't localise, and ipv6.google.co.uk doesn't resolve. It's pathetic, and comes under the "bad IPv6 setup/connectivity on the content provider's end" category. Google are the problem in my enjoyment of Google over IPv6 - not the ISP/tunnel (depending on where I'm at), and not transit. And yet now they have the audacity to tell the world how backward everyone else is.
And as for claiming Google "don't seem interested in actually deploying IPv6", that's just bullshit. Clearly they are, otherwise they wouldn't have bothered rolling out YouTube and all their other services over v6, nor would they have set up peering arrangements with HE, this ISP you cited, and I'm sure others. And they wouldn't be working on this whitelisting scheme now.
They have experimented with reduced feature IPv6 service provision, and now they're trying to set up a huge whitelist as a means to some other end - the fact that their ostensible aim is IPv6 deployment is overshadowed by their dominating, bureaucratic approach.
Dude, the rest of the Internet world is *already are being that cautious*! Don't you see? Content providers *already* refuse to deploy v6 because of these issues.
Content providers refuse to deploy IPv6 because they do not see themselves or their customers benefitting from it. Even if there weren't some sub-1% figure of customers who it was believed it would affect negatively, there'd still be no perceived benefit. It's unnecessary effort for no reward, and with risks far more significant than the problem of potentially pissing off your sub-1%.
But, once again, this is a chicken-and-egg problem. People won't build new, exciting applications on top of v6 technology if there's no one out there to use those damned applications in the first place.
So what's Google labs for? The A&A ISP managed to find enough people interested
I specifically remember the Dallas Morning News saying something to that effect 20 years ago, but I don't have any cites for that.
One local newspaper 20 years ago, and only if your memory serves you correctly: this is not evidence.
I've also heard that comparison many times, including recent New York Times analysis stating that they spend enough on distribution that they could give everyone with a subscription a free Kindle and save money in the deal.
That is a completely different statement altogether. All your quote is implying is that to give everyone a Kindle would involve fewer expenses - after some unstated time - than to continue physical distribution. Well, no shit, because the Kindle is a one-off cost followed by very small e-distribution costs. If 0 < a < c then na < b+nc for sufficiently large n, regardless of b. The assertion is trivial.
It's also worth noting the implication that/physical distribution costs are large/ - so large, in fact, that if the cost of distribution were covered entirely by subscription, it would for many newspapers imply that subscription is the largest revenue source. This runs precisely contrary to the assertion which started this thread.
You whine about ad hominems (when I didn't actually do any) and then commit one. Calling someone an idiot isn't an ad hominem.
Your ad hominem was to tell me that I'm not qualified to question because I haven't heard of some rule of thumb which you claim you read 20 years ago in some local rag. Your passionately built straw man is correct: calling someone an idiot isn't an ad hominem.
. That you can't understand the difference shows you shouldn't be speaking in public. Not that I'd restrict your free speech, but that you make those around you dumber just by opening your mouth.
You write very angrily. This is just a tech entertainment web site. It's not worth getting so worked up about something so minor. Relax, take a walk, have a beer, whatever. But please don't let my discussion with you stress you out so. Maybe you're not a troll - maybe you just get excited when your argument is challenged, and you express that excitement by degenerating your rhetoric with a level of rudeness. It's not the first time I've seen that on the 'net. No hard feelings, OK?
What about it? That video just illustrates precisely what this coalition is trying to deploy,
Sorry, what? The video illustrates how an ISP arrived at the stage of providing good native consumer IPv6. Yet Google, which has done no such thing for any consumer, thought his efforts not good enough when he tried to get whitelisted.
as Google has been doing this for a while now.
Google has done less for IPv6 than this guy's ISP has. They have:
1. Provided native IPv6 to all their clients, with a level of consumer IPv6 support unrivalled in the UK;
2. Got the biggest ISP wholesaler in the UK to fix their systems to make that possible, and to commit to future support of IPv6;
3. Ensured all mail, hosting, etc services provided by the ISP are IPv6-capable;
4. Gone around evangelising IPv6.
Again, is it annoying for the ISP? Yeah. I just don't care, that's all.
It's annoying for the ISP not only because it's a waste of time but because, as he commented, Google don't seem interested in actually deploying IPv6, and reject businesses such as his which are at the forefront of consumer IPv6 deployment. He did not speculate on what Google are actually up to, though I'm sure he has a much better idea than I do.
Frankly, I find it funny that guy doesn't understand why Google is being so cautious.
His business (and, as seems to the extent I've spoken to him, his passion) is in providing cutting edge Internet connectivity. He accepts the challenge of customers who are more demanding and less tolerant of failure than the Average User. Sometimes they fuck up - for example, they had fairly shitty custom written mail services for quite a while - but on IPv6, they have it just right. Google, in its control freakery, should not step so cautiously, and absolutely must not encourage the rest of the Internet world to do so.
Then again, he isn't in the business of making money based on eyeballs. Google, meanwhile, stands to lose real dollars if someone decides to, say, switch to Bing because Google's AAAA records cause
This comes back to the point I made ages ago that Google in its pedestrian businessman role (as an ad broker) is not appropriate as any sort of authority/organiser of the IPv6 transition.
If Google really wanted to make a difference, it would follow JANET's lead into properly documenting and cooperating with ISPs in implementing IPv6 multicast, and use its resources to showcase other IPv6 goodies. If Google really wanted to make a difference it could add AAAA records at some stage with a notice passing people to a legacy.google.com for people with fucked up IPv6 connectivity (which could be for any number of reasons/other than a nasty ISP-provided router/).
Let's make this quite clear: there is no clear advantage to the average consumer to making the average web site available on IPv6 rather than IPv4. Telling ISPs that, if they jump through more hoops than the best consumer IPv6 ISP in the UK had to, they might just get whitelisted by Google so that the consumer gets at best exactly the same experience as before, is a waste of everyone's time. "Being able to visit IPv6 versions of IPv4 web sites" is barely an advantage of IPv6. Adding some dumb hurdle to achieve it is of no use.
Anyway, if so many people avoid Google because of the broken router Abomination, a similar proportion will avoid this guy's ISP when his web site loads too slowly. They've taken the risk and used their extremely limited (vs Google) technical resources to identify and fix barriers to IPv6 deployment. Google are slow.
Slashdot Mirror
"stealing" is the act of depriving someone of the economic value of a thing.
When someone "pirates" your crappy software, you're not being deprived of the amount the market is willing to pay for your software.
Thus, by your definition, copyright infringement is not stealing.
Thank you for making a counterargument so concise. And remember, the market gets to determine the value of stuff you are trying to sell - not you.
The EeePc's 36W full load is quite similar to the C7's 55W, recalling that the latter has two 3.5" drives and various fans rather than SSD. If a casual browsing machine were what's required, they'd both be fit for purpose following de-moving-part-ification of the C7.
Others have mentioned laptops (including the Mac Mini, which is just half a laptop) as requiring less power. Well, yes - the aim isn't to find the lowest power machine with storage and net connectivity, as then my 4 AA cell powered ARM-based HP 50G calculator would shine. For it takes SD cards and has a TCP/IP stack written for it :-). Actually, the 8088-compatible Psion Series 3a would do better: it has a nicer keyboard than most palmtops and a screen which has lasted two decades without deterioration. Two AAs, *two* flash slots - albeit proprietary.
There is absolutely no way that you're reading 5W at the wall with one Core 2 Duo plus two spinning drives, sorry. If one of these is a 7200rpm+ desktop drive, 5W+ will be needed to spin one drive. I'm running 4 year old but stably performing WD SE16 drives, and I know I could do better, but not this well.
Which motherboard, drives, PSU? Is everything spinning?
I'm not surprised to hear that idle performance is excellent on a modern desktop CPU, and I've heard similar achievements with a single desktop HDD. The 45mm Core 2 Duos also idle very efficiently, but shoot up under load with no hardware acceleration to mitigate.
I know that I'm unlikely to get readings much below current unless I switch to a PicoPSU or similar, because inefficiencies in a desktop PSU shoot right up at very low power output.
Wotcha. I'm glad the thread's received quite a bit of interest!
(1) Why "should" disks be spun down on an idle server? It's not like CPU which can pretty much instantly switch between power modes: an HDD takes time to spin up and spin down, and its life is decreased in doing so. You may decide to spin down drives, but only after a sufficient period of idling. If you really want to save power, an idle server "should" be asleep with WakeOnLan enabled.
(2) The BeagleBoard has 128MB of what looks to be on-chip memory, which is of course going to be lower power than if you need to route to a separate connector and fit off-the-shelf DIMMs. But it is an unreasonably small amount for a server, which benefits from a large disk cache - especially when it needs to decrypt what it reads from the drives but can keep memory unencrypted.
(3) The 256MB NAND flash might be useful for a root filesystem (what's the wear-levelling algorithm on this thing?), but it's hopeless for server storage. It's true that flash is going to idle much more efficiently than an HDD, and it's one of the things I raised in OP.
(4) Could you give an example where the "Atom CPU when idle" is consuming more power than a BeagleBoard "system"? I'm not sure what sort of system you mean, but I'm obviously talking about a NAS.
That 50W is three orders of magnitude greater than 50mW is entirely irrelevant. The draw of the CPU/mobo isn't so dominant in a typical system, and you're comparing one comparatively low horsepower CPU on a low feature board with a more powerful CPU on a board with enough expandability to be applied as a small desktop/server.
The BeagleBoard is still cool, of course, but for other things. And if you compared power draw of a bare VIA mobo with the BeagleBoard, I'd expect the latter to win out.
Phrase play perceived pleasing? Perhaps.
Precious? Positively!
Pointless? Plainly.
Yes, I noticed that one straight after posting. Dammit :-).
For CPU cooling, I have replaced the stock PC2500e fan with a 40mm Rasurbo specced at 0.6W max, which I've then further slowed down. One of the dedicated case fans is used to cool the drives (quite effectively, so it'd be there regardless of CPU), and the other may be extraneous, as the PSU draws out air. It is left over from when I had a higher power board in there, particularly because I was considering a Zalman ZM-NB47J passive cooler for the C7 CPU and then keeping that case fan to ensure good airflow over it. This'd be quieter.
Many people seem to run completely motionless Via C7 boxes, including by modding this motherboard - although more likely is that you'd start off buying a fanless board. They're commonly available clocked at 1GHz, though 1.2-1.5GHz fanless seems possible.
All this said, I'd be interested in recommendations for ARM mobo/CPU combinations suitable for homebrew NAS/gateway applications, and to hear about crypto performance on these boards.
Are hp still emulating the saturn processor on their arm devices?
Yes, although some operations have been accelerated with a native rewrite. But the emulation allows a very mature product to be executed without the cost and bugs inevitable with a complete port. People don't use calculators for their horsepower, but for usability and portability. The HP RPN + CAS are very usable.
FWIW, I very quickly gave up learning Saturn assembler when I still had a real Saturn device :-). Today, there are well-known ways of escaping to native ARM.
Erm, no, for developers there have been Linux kernel crypto modules supporting the Via Padlock included since 2.6.11, and if you don't want them, you can always use the crypto instructions directly. The Java is just an API option provided by Via.
Erm, you're comparing apples and orchards.
45-50W is including UPS, PSU and everything spinning (and miscellaneous peripherals omitted from the above list, e.g. second Ethernet PCI, floppy, DVD drive, external modem), and with a discussion on how this is far from the lowest power solution and how I could improve things. The drives spin using around 8W each, for example.
The BeagleBoard draws 50mW at 5V DC with nothing connected, and would be hopeless for the NAS (and many other) applications described because it lacks the various ports from SATA to Gb NIC (or PCI for such expansion) which a higher power chipset provides, and lacks the horsepower necessary to make up for the lack of hardware crypto acceleration. USB NIC, now?
I had an ARM-powered primary desktop between 1990 and ~2000 (Acorn A3000, RISC PC), and now have an HP 50G calculator running off 4 AA batteries with a faster ARM CPU. I do appreciate the architecture and power efficiency of ARM designs.
Thanks for this (and to the other AC and Ogi_UnixNut).
And my wallet's set on you!
Sorry - although I mentioned that the Via's great for a home/office encrypted NAS, I perhaps wasn't clear that this is precisely the application I was talking about. I was just expanding the discussion on power-efficient CPU applications, and implying that, when considering energy efficiency, a low power CPU with dedicated circuitry for popular complex operations might be the way forward.
Yes. The bottleneck for ssh probably becomes the HMAC - this thread is enlightening. SHA HMAC is afaict considered difficult to accelerate with Via's implementation, though a patch may exist. With a stock (AES-accelerated) build, software MD5 is quicker.
Wait, what? I was talking about a specific application in the general topic of power-efficient CPUs. I've yet to see a comparison of current high end Intel vs Via, e.g. Atom D510 vs Via Nano. If you genuinely need to sort all day, you probably have the intelligence and resources to prepare yourself, so you might be better off building a hardware sorting algorithm in an FPGA. You're unlikely to cluster low-power CPUs on off-the-shelf motherboards because your interconnect will be shit.
Also, what is the 1GHz referring to? Modern Via chips typically run up to 2GHz, and are architected differently from Atoms, so a straight MHz comparison is (as always) meaningless. For example, the Atom is (afaict) still an in-order CPU.
I'd gladly pay up...
The Via C7/Nano seems to be a great chip for a home/small office server, what with its built-in AES encryption making it faster than even a high end Xeon without hardware acceleration. My current setup consists of 2*WD SE16 hard drives, APC UPS, 80+ Corsair PSU, PC2500e Nano mobo with 1GB, and a couple of 80mm case fans, together running under 50W idle, and only 7W more at full CPU load. If I were to replace the Corsair with a fanless PSU good up to 80-120W I might get an extra 5-10% efficiency; I could wipe out the case fans probably with no problem (2-3W, say), especially if I replaced hard with spinning solid state storage, and that of course would shave off around 15W. Substitute a large fanless heatsink for another W (or just get a fanless motherboard/CPU in the first place). But even as-is, it's a good improvement on my previous regular desktop CPU-based setup.
For something which is on 24 hours a day, going several months between reboots and stressed only in the IO and encryption departments, I see no reason to use a full-power desktop processor. So, what problems have you guys encountered which has meant you haven't ended up with this option?
...
Honestly, I've never understood why people, when faced with an idea they don't like or can't understand, just fall back to deriding the source of the idea.
I dislike the idea, *and* I don't trust the source of the idea. These two intertwined issues. Even if the idea was proposed by an entity I felt confident dealing with, I still wouldn't like the idea. Even when Google pushes good ideas, I still don't want Google implementing those ideas if they affect others on too large a scale because I don't feel confident dealing with Google. But, in this case, we have a bad idea from an entity which would be expected to propose this sort of solution (because it fits in with its modus operandi) rather than something better for the Internet as a whole.
It is perfectly legitimate to attack both the idea and its proposer's motives, just as people are free to attack both the idea/implementation of Obama's healthcare plan and question his motives. For motives will reveal any gotchas and affect the implementer's future direction. No action is performed in an altruistic vacuum, and you're regrettably naive if you think Google's final aim is "bringing IPv6 to the masses". Its final aim is to make money, and if we can't clearly see how it's doing that by pushing IPv6 as it is, it's because we lack the knowledge, not because Google are altruistic. Who knows, maybe they're negotiating with big ISPs to do something fancy with multicast? Either way, I'd like to know if they expect me to cooperate.
As a parting note, I want you to consider the difference between reasonable and Google's method in encouraging the casual enthusiast to adopt IPv6 (and every technological take-off begins with the casual enthusiast).
Expected:
1. Set up IPv6 tunnel, receive feature parity Google service via IPv6.
Experience:
1. Set up IPv6 tunnel;
2. Be surprised that you're not getting any Google IPv6 connectivity;
3. Find out that www.google.com isn't providing AAAA records;
4. Check with your tunnel provider, and find that they have signed up to some Google whitelist;
5. Assume that this means you have to make DNS requests from an IP address in their range, so get your local forwarding DNS server working on IPv6;
6. Find out that www.google.com isn't providing AAAA records;
7. Read up further, and find that the whitelist is actually a specific list of their DNS servers;
8. Find out what these DNS servers are, which involves a dig to list all the AAAA records for the local list of five or six caching DNS servers;
9. Add all the records to bind;
10. Find that you're finally getting IPv6 addresses (woohoo!), but that resolution is intermittently slow;
11. Ping each DNS server to find out which is lowest latency - they're all at least double the latency of your ISP's local DNS server, but whittle down to two;
12. After more experimentation, find that one of these servers takes around 5 seconds to respond to half the requests;
13. Whittle down to one server of reasonable speed;
(14. For end-to-end IPv6, send IPv6 address for your local DNS server via RDNSS, the recent extension to the router advertisement protocol, but find that Windows 7 ignores this extension entirely. Not a great problem for me, but a barrier to those who might want to advertise the IPv6-only DNS server found in 13 directly, as they'll have to set up DHCPv6. And, if you have XP, you're SoL.)
The ironic thing here is that this describes an experience with a service not only whitelisted by Google but with administrative input from Google employees. If Google wasn't such a control freak, the user would be done at step (1).
Ah, engineers. You have the brainpower to build a Rubik's Cube solving robot, but you skip the processing which concludes that dark green on dark blue makes for an illegible web site.
(I know that sounds like a troll, but I'm interested to understand why some people with stellar abilities will nevertheless end up doing something apparently trivial like not checking that their work is legible.)
Thinking about it, since Microsoft has been onboard with IPv6 before anyone had even heard of Google, they'd be better poised to ensure a smooth transition by pushing an update which regularly checks for good IPv6 connectivity and temporarily disables attempts to resolve AAAA records if such a check fails. A simple registry/UI change should disable this service, of course.
As for where the servers are which respond to the connectivity checks - well, MS can host them and rejoice in the data it collects by hearing the IPv6 pings. They already get to phone home in their EULA for antipiracy and update checks, and the world seems to be happier with these potentially much greater intrusions.
The Linux community could produce a similar daemon. Hell, cooperate with Microsoft so you can use their ping collectors if you want, as otherwise angry rival newspaper publishers cooperate on technologies for the common good such as environmental measures. However, already having the resources to provide updates, I'm sure the Linux community can muster up the bandwidth to hear a ping every so often.
Again, the ISPs have made this bed. This guy may be one of the shining lights, but he's in an industry that's just barely competent enough to keep a regular v4 network up and alive. You can hardly blame Google for doing their diligence.
Their diligence? Google are irrelevant when it comes to pioneering either development or deployment of network layer technology. Yet instead you're telling off ISPs for not somehow not providing IPv4 properly, when they've managed to help grow the consumer Internet in the past 10 years at a rate unrivalled by probably any other consumer tech?
Google are FUDing, because it's in Google's interest to spread FUD as a precursor to the "but if you just put all the information/control in our hands, we'll make it all better..." which forms the basis of every single business decision they make.
Oh come on, quit being so dramatic. Google never rejected anyone. Did they make those guys go through a bunch of hoops? Yes, they did. But the guy flat out stated that he expects to close a peering agreement with Google.
This is because he is stubborn and persistent. Most ISPs would just stop caring before they'd even started.
It's gonna be a concerted effort between ISPs *and* content providers, and this scheme finally has a chance of getting content providers onboard.
All it does is give content providers a chance to reduce the inconvenience for under 0.8% of users once or twice (until they figure out the problem or choose the legacy URL) in the event that providers believe simply adding AAAA records is correct migration procedure. Since adding AAAA records with the current reliability of global IPv6 connectivity is a shit idea anyway - i.e. the problem is as likely to be bad IPv6 setup/connectivity on the content provider's end, or somewhere in the middle - you're trying to solve a problem by an initiative which implies a poor method.
Indeed, Google still doesn't provide usable IPv6 search for anyone outside the US, because its ipv6.google.com won't localise, and ipv6.google.co.uk doesn't resolve. It's pathetic, and comes under the "bad IPv6 setup/connectivity on the content provider's end" category. Google are the problem in my enjoyment of Google over IPv6 - not the ISP/tunnel (depending on where I'm at), and not transit. And yet now they have the audacity to tell the world how backward everyone else is.
And as for claiming Google "don't seem interested in actually deploying IPv6", that's just bullshit. Clearly they are, otherwise they wouldn't have bothered rolling out YouTube and all their other services over v6, nor would they have set up peering arrangements with HE, this ISP you cited, and I'm sure others. And they wouldn't be working on this whitelisting scheme now.
They have experimented with reduced feature IPv6 service provision, and now they're trying to set up a huge whitelist as a means to some other end - the fact that their ostensible aim is IPv6 deployment is overshadowed by their dominating, bureaucratic approach.
Dude, the rest of the Internet world is *already are being that cautious*! Don't you see? Content providers *already* refuse to deploy v6 because of these issues.
Content providers refuse to deploy IPv6 because they do not see themselves or their customers benefitting from it. Even if there weren't some sub-1% figure of customers who it was believed it would affect negatively, there'd still be no perceived benefit. It's unnecessary effort for no reward, and with risks far more significant than the problem of potentially pissing off your sub-1%.
But, once again, this is a chicken-and-egg problem. People won't build new, exciting applications on top of v6 technology if there's no one out there to use those damned applications in the first place.
So what's Google labs for? The A&A ISP managed to find enough people interested
I specifically remember the Dallas Morning News saying something to that effect 20 years ago, but I don't have any cites for that.
One local newspaper 20 years ago, and only if your memory serves you correctly: this is not evidence.
I've also heard that comparison many times, including recent New York Times analysis stating that they spend enough on distribution that they could give everyone with a subscription a free Kindle and save money in the deal.
That is a completely different statement altogether. All your quote is implying is that to give everyone a Kindle would involve fewer expenses - after some unstated time - than to continue physical distribution. Well, no shit, because the Kindle is a one-off cost followed by very small e-distribution costs. If 0 < a < c then na < b+nc for sufficiently large n, regardless of b. The assertion is trivial.
It's also worth noting the implication that /physical distribution costs are large/ - so large, in fact, that if the cost of distribution were covered entirely by subscription, it would for many newspapers imply that subscription is the largest revenue source. This runs precisely contrary to the assertion which started this thread.
You whine about ad hominems (when I didn't actually do any) and then commit one. Calling someone an idiot isn't an ad hominem.
Your ad hominem was to tell me that I'm not qualified to question because I haven't heard of some rule of thumb which you claim you read 20 years ago in some local rag. Your passionately built straw man is correct: calling someone an idiot isn't an ad hominem.
. That you can't understand the difference shows you shouldn't be speaking in public. Not that I'd restrict your free speech, but that you make those around you dumber just by opening your mouth.
You write very angrily. This is just a tech entertainment web site. It's not worth getting so worked up about something so minor. Relax, take a walk, have a beer, whatever. But please don't let my discussion with you stress you out so. Maybe you're not a troll - maybe you just get excited when your argument is challenged, and you express that excitement by degenerating your rhetoric with a level of rudeness. It's not the first time I've seen that on the 'net. No hard feelings, OK?
What about it? That video just illustrates precisely what this coalition is trying to deploy,
Sorry, what? The video illustrates how an ISP arrived at the stage of providing good native consumer IPv6. Yet Google, which has done no such thing for any consumer, thought his efforts not good enough when he tried to get whitelisted.
as Google has been doing this for a while now.
Google has done less for IPv6 than this guy's ISP has. They have:
1. Provided native IPv6 to all their clients, with a level of consumer IPv6 support unrivalled in the UK;
2. Got the biggest ISP wholesaler in the UK to fix their systems to make that possible, and to commit to future support of IPv6;
3. Ensured all mail, hosting, etc services provided by the ISP are IPv6-capable;
4. Gone around evangelising IPv6.
Again, is it annoying for the ISP? Yeah. I just don't care, that's all.
It's annoying for the ISP not only because it's a waste of time but because, as he commented, Google don't seem interested in actually deploying IPv6, and reject businesses such as his which are at the forefront of consumer IPv6 deployment. He did not speculate on what Google are actually up to, though I'm sure he has a much better idea than I do.
Frankly, I find it funny that guy doesn't understand why Google is being so cautious.
His business (and, as seems to the extent I've spoken to him, his passion) is in providing cutting edge Internet connectivity. He accepts the challenge of customers who are more demanding and less tolerant of failure than the Average User. Sometimes they fuck up - for example, they had fairly shitty custom written mail services for quite a while - but on IPv6, they have it just right. Google, in its control freakery, should not step so cautiously, and absolutely must not encourage the rest of the Internet world to do so.
Then again, he isn't in the business of making money based on eyeballs. Google, meanwhile, stands to lose real dollars if someone decides to, say, switch to Bing because Google's AAAA records cause
This comes back to the point I made ages ago that Google in its pedestrian businessman role (as an ad broker) is not appropriate as any sort of authority/organiser of the IPv6 transition.
If Google really wanted to make a difference, it would follow JANET's lead into properly documenting and cooperating with ISPs in implementing IPv6 multicast, and use its resources to showcase other IPv6 goodies. If Google really wanted to make a difference it could add AAAA records at some stage with a notice passing people to a legacy.google.com for people with fucked up IPv6 connectivity (which could be for any number of reasons /other than a nasty ISP-provided router/).
Let's make this quite clear: there is no clear advantage to the average consumer to making the average web site available on IPv6 rather than IPv4. Telling ISPs that, if they jump through more hoops than the best consumer IPv6 ISP in the UK had to, they might just get whitelisted by Google so that the consumer gets at best exactly the same experience as before, is a waste of everyone's time. "Being able to visit IPv6 versions of IPv4 web sites" is barely an advantage of IPv6. Adding some dumb hurdle to achieve it is of no use.
Anyway, if so many people avoid Google because of the broken router Abomination, a similar proportion will avoid this guy's ISP when his web site loads too slowly. They've taken the risk and used their extremely limited (vs Google) technical resources to identify and fix barriers to IPv6 deployment. Google are slow.