tbake0155

I'm not sure I follow you. 70A at 1.5V would produce 105W, completely reasonable power draw for a modern cpu. It's exactly because cpus are using such low voltage that high current is needed. With any power delivery mosfet, you are better offer going beefier than needed because the efficiency will be better and it will produce less heat (effectively efficient and low heat output are synonymous in this case). Ideally, engineers can maximize efficiency but it's not always up to just engineers, the bottom line comes into play as well.

colonel_mortis is essentially correct. If you wanted to have all of the resources appear as one computer, you would have to run some type of simulation layer on either a 'master' machine or one of the workers in the cluster that would then behave as a master and a worker (probably not ideal). This would be an OS-like piece of software, and would most likely severely hinder performance, which is why you probably won't find something like that available, but in theory you could write something up yourself (it would be a huge undertaking). In general, leveraging a cluster requires you to write multi-threaded applications that can use the resources of several machines, usually by working individually and then combining the results when all of the workers are done. This means you need an algorithm that can be tackled with multiple threads simultaneously without the threads depending on each other (or at least minimally so, as they could use shared memory and access the same variable, but not at the exact same time due to locking), which is pretty normal in scientific computation but not necessarily for all cases. There is cluster software available like Slurm that you can install to manage your cluster more easily. Whatever you do, you will need to use a library like MPI when writing your code so that your nodes can work together. Take a look at this: https://shop.pimoroni.com/products/cluster-hat You might go through the instructions for setting that up to get an idea of how to proceed. You could probably even use the modified raspian image provided by the manufacturer, but you would have to make some changes since the instructions are for setting up USB network connections rather than ethernet.

I came to say this. Not all passwords are hashed using the same algorithm. Plus, pre-computing rainbow tables for 8 characters is quite feasible, so 8 character long passwords have been pretty insecure for a while.

I would argue it has legitimate use cases but it also has shortcomings, as you mentioned. In rural areas it will not make much sense but in urban areas, where having repeaters in close proximity will generate coverage for a lot more people and for a lot more use cases, it can be effectively implemented.

Speed isn't the only change that will come with 5G, the big change with 5G is latency. Since the wavelengths are so much shorter in the higher frequency waves that 5G uses, the latency for consumer wireless connections will be drastically reduced, in the order of possibly a 50 times reduction in latency from 4G.

An sftp server would be a good use for your raspberry pi, as long as you don't expect a ton of traffic simultaneously. I think if you use a raspberry pi, you should consider a usb hard drive, since SD read/write isnt great, and freeing up the SD for the operating system will almost certainly give you better performance, especially if you have simultaneous users. Linux will also allow you to create software raid with USB drives, which could further improve performance. I would recommend you try the raspberry pi, it will be a fun project if nothing else. If you're moving large film files, it might be too slow but since you have one already Id say try it. If you follow a few safety measures I think you will be fine. Forward the ports you need for sftp through your router, and only give the sftp software access to those ports in your software firewall on the machine running the sftp server. Use strong passwords, and ideally dont use the administrator account to run the server (you should not use it as a best practice anyway). If you get concerned, consider figuring out how to log internet traffic on the open port(s) and make sure the traffic you see is you or your users. You will see the IP address , which you can confirm with your users. Even if a connection attempt fails, you will see it. If you see a lot of strange traffic, consider taking measures. Generally though even if you saw a strange connection attempt, if your passwords are secure you should be fine even getting brute forced. I'll briefly explain how port sniffing would work. Imagine a bot attempts to get a response from your IP address at every single port, this is simple enough using a tool like netcat. This wont really be the approach the bot would take though, unless they knew you were a high value target. Since most people with weak security will use default ports, a smarter bot might only check certain commonly used ports (ssh, http, https). That will allow the bot to scan exponentially more IPs in a given time frame, and so the bot has higher odds of finding a target. But, for the sake of the example, lets imagine a bot will sniff every port. Using netcat, you can tell the difference between an open port with restricted access and a port that is complety shut off. If your router is forwarding no ports, there will be no 'smell' and the bot moves on. If your router is forwarding ports, that port will give itself away to a program like netcat by rejecting the connection (the foundations of the internet do not, and did not from the beginning, focus on security, it was left up to software developers to implement). Now the bot can try to find a way in. I do not know exactly how a hacker would operate from there, but they would have time on their side. Plenty of systems are open to the internet so its not intrinsically a bad idea. However most servers are not running on bare hardware and they have security features in place. But systems that are vulnerable and get targeted are sometimes breached.

I apologize for the delay in responding. Generally, sftp runs through the default ssh port. However, it sounds like you are running the sftp server from a Windows machine, and if that's the case, I would have to get some more details about the sftp server software you are using. I initially wrote my response assuming you configured sftp yourself, but if you are using software you will have to tell me what the software is. All of my experience using ssh/sftp was with Linux machines or using Git Bash, Cygwin, Putty, etc, to log into Linux machines. On a Linux machine, ssh/sftp runs by default on port 22, but that can be configured. Moving it from port 22 is probably a good idea, since most bots are sniffing for particular, widely used ports. If you pick a random port over 1200 or so, you'll be less likely to get sniffed. If your software is not using port 22, that's good. Here's the reality though: If your password is strong enough, you should be ok. By strong enough, I mean use 20+ characters with caps, lower case, numbers and special chars. That will take eons to crack currently. Do not give administrator privileges to any other users except yourself, and encourage or enforce strong passwords for your users. Better yet, do not use an administrator account to run the server, so that if someone is able to log in to your account, they will still not have admin privileges. You will need to secure your admin account with another strong password. The issue with having an ssh port open is that unless you are monitoring it, someone can attempt to break in by trying every password combination. This would take a ridiculous amount of time, especially if your password is strong. However, they can perform this attack from anywhere in the world that has internet, which makes this a vulnerable attack vector if you were to say leave the port open for years and years without monitoring it. Regarding dynamic IP, that should help if someone was in the middle of brute forcing you for a really long time, but I doubt your IP rotates frequently enough that it will make much difference. If your password is strong enough, your IP could not rotate for years and it would not really matter. I would not consider a dynamic IP as a security measure and I would not rely on it. Regarding a virtual machine, that would still be my advice for you. Install VirtualBox and then create a VM using Ubuntu or Windows. Set up the VM so that it has an internet connection and so that you can access it from other computers on the local network, then forward from your router to the VM instead of your base Windows OS. If someone got access to the VM, they would be stuck in there. If all that is running in your VM is the sftp server, then that is the only thing they would have access to. It's also the case that an intruder will not be able to interact directly with the hardware on your machine, which is the best practice for security. If you have more questions feel free to ask!

What you want to do is most definitely possible, the easiest way will be php scripting on the server side. It's not a trivial task and I have no idea if there is software available that you can use out-of-the-box to achieve this. For someone well versed in web design, this is relatively straight forward (although would take a considerable amount of time as it is non-trivial). However, I would offer you a few pieces of advice. Is your sftp server running inside a virtual machine? If it's not, it probably should be. Ideally, even if the server was running on dedicated hardware, it would be a best practice to put the sftp server in a virtual machine instead of running it directly on the hardware. Second, It's generally not a good idea at all to open up your home internet connection to the outside. Especially opening well-known types of connections, such as ssh (and sftp). If you are going to open up ports on your home system, you should really setup a powerful router (i.e., build a machine and run pfsense, not buying a router). There are bots sniffing for port 22 open on EVERY possible IP combination, all day long, every day, because why not. You can verify this if you open up port 22 and log connection attempts, I've done this and I had numerous attempted connections every day. So if you open ssh to the outside world, take some extra precautions. You are creating a giant attack vector into your home system by not only opening it up to the outside internet, but especially by running the sftp server on (presumably) your desktop machine. The safest thing to do would be pay for a virtual server from Digital Ocean or AWS, however, storage is costly with those services so I can understand you not wanting to do that. If you are in university, there's a good chance you will have the ability to host a website on the school's servers (you might look into it), which would also very likely be using massive storage systems so you would have plenty of memory. I know having an sftp server running is cool, but it's also something that you should do with consideration to the potential vulnerabilities you will create. Sorry for not being more helpful in regards to original question, but hopefully this will give you some insight into best practices regarding running an sftp server from your home machine. If anyone thinks I am way off base then please correct me!

I'm truly not informed enough to support what I'm saying, but I would not intentionally spread false information. I am certain that I previously saw evidence of what I'm referring to, and thus did not create the idea from my imagination. Also, when I say not visible on a voltmeter, I mean a multi-meter tool and not an oscilloscope. You would certainly be able to see these types of spikes on an oscilloscope. Although I cannot argue with your stance, I have found more evidence that what I'm referring to does in fact exist. It's my (limited) understanding that even the discharge of each spark plug will cause transient voltage spikes in the system (because it's a load that appears and disappears). Here's an article that I found: http://www.electronicdesign.com/power/eliminate-those-automotive-load-dump-circuit-protection-headaches. It appears that automotive power supplies are inherently noisy and transient. I will concede to your expertise, and it's possible that I'm mistaken and not interpreting the materials I'm presenting to you correctly.

Transient Over-voltage in a vehicle http://m.littelfuse.com/~/media/electronics_technical/application_notes/varistors/littelfuse_suppression_of_transients_in_an_automotive_environment_application_note.pdf Sorry for the double post, but searching for transient over-voltage pretty much clears this up. If you still disagree, I consider you take your argument to an electrical engineer and not debate with me, since all I can do is regurgitate what I've seen on this topic. Take care.

I'm 99% sure that it's not false. Unfortunately I can't seem to find a resource to describe this (searching google for car alternators is a huge hit). The spikes will be fractions of a second and will not show up on a voltmeter, but they will destroy sensitive electronics. If I had more time I'd search for it, but we can agree to disagree since it's not really a major issue. However, did you read the rest of my post? There are power distribution boards available designed specifically to let a person power a motherboard from a car battery, although they can certainly be adapted for other purposes. BTW, directly from the information listed with the power board I linked to in the previous post : "M4-ATX-HV is a wide input (6 to 34V) vehicle / car / battery operated<br>ATX power supply capable of surviving tough car engine cranks (down to<br>6V) as well as transient over-voltage situations" Regardless of the spikes, anyone can hook a motherboard up to a car's battery and create a mobile computer if that is what they intend to do.

You can power a x86 motherboard from a car's battery/alternator. Even though the motherboard needs 12V on the highest volt rail, and most car batteries are around 13 V when charged, the alternator will create spikes of over 100V (you can't see this on most voltmeter's because it happens too fast). Those votlage spikes would kill a computer connected directly to the battery. Of course, who would wire it that way. I was only pointing that out because I think it's interesting. There are power distribution boards made for exactly this. You can hack a board like this out of some vehicles (since all modern cars have ECU's). Here's an example of what I'm talking about: https://www.amazon.com/M4-ATX-HV-Intelligent-Power-Supply-6-34V/dp/B005TWE4YM I've personally used that board with a 13 Volt power source and was able to run an underclocked i7 6700K (2.0 Ghz, approx. 30 Watts at full load) and a GTX 750 ti. There's a lot of cool stuff you can imagine doing with a mobile computer like this. BTW, although a car will let you keep the computer on indefinitely as long as the car's motor is running, you can also build a mobile computer with a battery using a board like this... Endless possibilities!!

While it is true that TDP is not the power used by the CPU, it is also true that all of the power that goes into the CPU is eventually released as heat. So although TDP != power draw, we can simplify and say that TDP == power draw. That's why it is used as a simplification for the masses, who generally don't care about having an expert's understanding of TDP.

1.62 volts is way too high. The 9590 is just a binned and overclocked 8350, but the reality is that they don't generally overclock more than the later 8350 chips anyway. I have an 8350 with a custom water loop cooling it. 3 radiators and an EK water block. I was originally going to water cool the GPU with this set-up also. It's way overkill for just the CPU. 5.1 GHz is possible, I've done it and I've even done some serious stress testing but it's not 24/7 stable. I never used more than ~1.575 V either. Anything above ~1.51 and the heat starts pouring out. This is on a sabertooth 990fx. And as others have mentioned these CPUs should not be operated at temperatures above 70 degrees celcius. Personally I aim for 65 max. The CPUs will also hard thermal throttle. So while you might boot at 5.1 GHz, if you actually look at your cores using HWmonitor, for example, you will see that some of your cores are dropping to zero percent utilization during a (real) stress test. I can almost guarantee that you are thermal throttling unless your temperatures are wrong, but with that voltage I'm sure it's right. My 8350 is really tough. During testing to find it's limits it rarely froze the computer, it would just start dropping off cores until it can manage the temperatures.

Dendrite formation is still a problem as far as I know. The effects have been largely mitigated by using a carbon (graphite) anode. Ideally, the cathode and anode would be lithium. However, the problem is that dendrites form on the anode during charging and lithium prolifically produces dendrites during charging. (I was incorrect earlier stating they project from the cathode.) There's some research into the possibility that the formation of dendrites on the anode is from contaminants below the surface of the anode, and not specifically from the material itself, but I believe that study is still underway. So modern li-ion batteries and even li-po batteries have dendrite formation, it's just that the amount is typically not enough to cause damage because the anode is graphite. I'm beyond my expertise in this area, so I will concede from here!