Trippling internet speed with pare of AA batteries, ethernet cable and electric tape.

[EasyGoing1]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

Unless Godaddy is actually hosting your IP address range via VPN or other routing chicanery, the reverse DNS lookup request necessarily goes back to your ISP, unlike the forward DNS which goes to whoever is hosting your domainname. Reverse DNS related to ISP block lookup, it needs to start with the frontmost digits as those are the most significant (rather than the lastmost name in forward lookup as that's the domain name).

You're absolutely right on that ... I responded to that just from seeing DNS and ISP ... my bad ... I stand corrected.

[eccerr0r]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

Unfortunately with big ISPs they have a bureaucracy going on there, and nobody wants to do things... They did set up a website to do these RDNS setup but it constantly breaks (and the authentication...) but they did not put up IPV6 setup. Plus it's still not clear because though effectively I have that /64 I don't think they really want me setting up that many RDNS records though theoretically I should...

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

...my ISP has the $ORIGIN/SOA for the reverse as they should...

Is there no way to work with your ISP to try and show them how to fix the RDNS stuff? It sounds like you know a good deal about networking, more than I do, that's for sure. In the past, I've worked with some companies, after finding problems, and gave solutions on how to fix them. Some companies implement the fix, some just don't care.

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

YES IT IS ... and thats where the issuing authority needs to say "look ... you don't need all those addresses ... learn some basic networking principles and manage your assets accordingly" ...

Assigning public ip addresses to workstations is both a security risk, and a pig ignorant way to run a network.

I brought it up in the meeting after I became manager. They didn't know and the head IT guy in charge at the time was asked why. With FedEx, for whatever reason, they had to be on public. We weren't allowed to touch them. I mean, we could repair them, etc, but they had special stuff on them that fedex used. I don't know much about how that was setup.

We spent the night switching all the others over to private.

I think moving to IPv6, right now, is the only viable solution. I think even if people were forced to manage their IPv4 addresses a bit more wisely and give back the IPs they weren't using, it's just prolonging the inevitable. Eventually, even if we get the IPv4 addresses back, we'll run out again.

[eccerr0r]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

Oh I'm quite familiar with subnetting .. And very vew corporations would subnet their public addresses ... what would be the point? Their Internet connection terminates at a router then at a firewall but at the router, natting can start, so breaking up their assigned network addresses makes no sense really. If they have multiple locations, the ISP will assign their own subnetted blocks of addresses to each location. So if anyone is sucking down broadcast and network addresses, it would be ISPs ...

Who could have predicted that the ISP now should be delegates of IP addresses and not companies? Before when the Internet was small, this was the "best" way and have every machine visible on the network... now we know this is very bad.

I've often wondered what class of addresses the main authority assigns to ISPs and what not ... Ive only ever worked with 24 bit masks have only gone higher than that in the private address space.

IANA chooses depending on need and how much they pay...

Thats a curious statement ... what do you mean by this?

As in the SSL thread I expect network pipes to be a pipe. And when that is true, I have a machine on a public static IP, no third party entity can choose to remove my posting because someone feels like it. I can advertise all I want on my IP address. It's mine. In a facef**k posting you're restricted to what facef**k lets you do.

The issue of reverse DNS should be with the company who hosts your DNS record ... Godaddy in my case. The ISP has nothing to do with it, unless their DNS servers are also the root servers to your domain name. If so, just switch to someone else for record hosting. I think google does it for free ... you still pay whatever company you pay now for the domain name, but you go into your record and set the IP of whatever name server you are going to use and then all hits to your domain will start with the new name server... and I'm sure I can say with confidence that google has RDNS working with IPV6.

To find out who hosts your record you can go to a command prompt and do an nslookup against your domain name with a soa query type.

Unless Godaddy is actually hosting your IP address range via VPN or other routing chicanery, the reverse DNS lookup request necessarily goes back to your ISP, unlike the forward DNS which goes to whoever is hosting your domainname. Reverse DNS related to ISP block lookup, it needs to start with the frontmost digits as those are the most significant (rather than the lastmost name in forward lookup as that's the domain name).

Say the forward badcaps.net -> it looks for a DNS server that handles .net, then it looks for badcaps and finds 168.235.77.64.

Reverse DNS actually uses the same forward DNS system to simplify things, but uses a different hierarchical path. Because the first number is the most significant (looking up from the 64 really has no meaning because there are 16777216 possible reverses that at best would have to be searched as a huge list), it actually looks up a virtual hostname "64.77.235.168.in-addr.arpa." The in-addr.arpa delegate DNS host can't possibly know how to know what every mapping is, so it looks for an origin match and delegates pieces back to servers that would likely know. Since ISPs tend to have whole /24's, /16's, etc. they are the likely candidate to handle RDNS lookups because you can't have two different people $ORIGIN/SOA for 235.168.in-addr.arpa. Only one can be delegated, and that ISP where packets route to 168.235.x.x gets routed to is the best choice (Second best choice would be the upstream provider). This is why despite my top level forward DNS hostname is resolved by GoDaddy, they cannot also fix my reverse DNS - my ISP has the $ORIGIN/SOA for the reverse as they should. It reduces the amount of collating of data at a central point, and instead allows hierarchical lookup. And when the ISP's reverse lookup server is dead, just that ISP's reverse lookups fail.

As an aside, GoDaddy was willing to let me put in an NS record so I can manage my own third level, so my private DNS server I can set up an unlimited number of my own forward lookup hostnames as I choose without needing to contact GoDaddy...

[EasyGoing1]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

I think one of the problems with IPv4 was like in Deposit. The PCs had public IPs for whatever reason. We did business with banks and FedEx, but mostly big banks. Banks that small banks do business with (ie, Juniper). Class C was too small, Class B gave us waaayyy more IP addresses than we'd ever use. But I don't think there was away to get a partial class B network. I don't think they could have asked the US registrar's for 1,000 class B public IP addresses, for instance. I think you need to get the full 65,536 addresses. That's a lot of wasted IPv4 addresses!

YES IT IS ... and thats where the issuing authority needs to say "look ... you don't need all those addresses ... learn some basic networking principles and manage your assets accordingly" ...

Assigning public ip addresses to workstations is both a security risk, and a pig ignorant way to run a network.

[EasyGoing1]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

Even if companies had to give back the IPv4 addresses they weren't using, that's still a temporary solution. Eventually, we'll be back to where we are now, with running out of IPv4 addresses. It might take a good amount of time, but I think we just need to sit back down at the drawing board and either revise IPv6 or, like you suggested earlier, come up with a better protocol all together (I'm thinking something like Spork/IP, that sounds good!)

I'm HOPING that a new protocol would emerge by then that would solve two problems ... address space and packet overhead which is disgusting in TCP/IP.

[EasyGoing1]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

There are a lot of reserved addresses in ipv4 so the 4 billion number is not accurate:
- Network and broadcast addresses. Ultimately every time a network is broken down to a smaller chunk, more addresses are 'wasted' to segregate traffic from each piece.

Oh I'm quite familiar with subnetting .. And very vew corporations would subnet their public addresses ... what would be the point? Their Internet connection terminates at a router then at a firewall but at the router, natting can start, so breaking up their assigned network addresses makes no sense really. If they have multiple locations, the ISP will assign their own subnetted blocks of addresses to each location. So if anyone is sucking down broadcast and network addresses, it would be ISPs ...

Normally a class C network subnet would only have 2 wasted addresses (broadcast and network address)

I've often wondered what class of addresses the main authority assigns to ISPs and what not ... Ive only ever worked with 24 bit masks have only gone higher than that in the private address space.

I chose to have static IP so my voice can be heard.

Thats a curious statement ... what do you mean by this?

I have a dual stack IPV4 and IPV6 setup through 6RD, but unfortunately there are problems with IPV6. There are a lot of people who don't have IPV6 setup properly, and I can't even set up my IPV6 properly because my ISP can't set up IPV6 reverse DNS properly. As far as I know I am fully functional listening to IPV6 requests but since reverse DNS does not work, people reject connecting to my machine... Figures, have to use IPV4 which reverse DNS does work.

The issue of reverse DNS should be with the company who hosts your DNS record ... Godaddy in my case. The ISP has nothing to do with it, unless their DNS servers are also the root servers to your domain name. If so, just switch to someone else for record hosting. I think google does it for free ... you still pay whatever company you pay now for the domain name, but you go into your record and set the IP of whatever name server you are going to use and then all hits to your domain will start with the new name server... and I'm sure I can say with confidence that google has RDNS working with IPV6.

To find out who hosts your record you can go to a command prompt and do an nslookup against your domain name with a soa query type.

And yeah, I virtually have a /64 or something like that IPV6 chunk... No way I can use that many addresses...

The number of possible addresses in v6 is mind boggling...

[Per Hansson]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

IPv4, been like that for a long time now due to how segmented the net have become.

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

...And it creates also another issue which is routing related: before you could look at an IP and know which country it belonged to.
Not anymore, the address space is so segmented that addresses very close in numbers can be on different continents.
This makes routing tables very big and complicated.

You're saying with IPv4 addresses, you cannot tell which country the address belongs to now? Or are you saying that's with IPv6 addresses?

Thanks.

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

- A class A is a /8, a class B is a /16, and a class C is /24. As seen in my case I have a /29 partial C so it is possible to break a network outside of an A, B, or C class. However every break, as said, will have more overhead addresses wasting more addresses, so it's a tradeoff (plus the breaks need to be an integer power of 2). To better save IPV4 addresses everyone needs to be on the same and huge subnet, and people know what messes comcast has to deal with.

I know this as subnetting. My understanding was the main reason you'd subnet a network was to cut down on collisions. Are there other reasons?

...- There is indeed more chatter with IPV6, but that's because it's autoconfigure plus the additional bytes for specifying the addresses - and it's only on the local segment. But still, the amount of chatter packets is not that large - it's mainly the router advertising itself to all on that segment. The endpoint machines do not need to keep chatter the router unless it has real data it needs to route out.

I've done things on one of my servers to try and cut down on some of the chatter. Stuff with IPv6 ICMP packets (I can check into what I did, if you really care). I did this at a firewall level, to try and prevent abuse. It's good to know the amount of chatter is not that large.

- UDP does not mean faster than TCP. Both are running over IP so there's always that limit, but the main difference is that TCP holds state and "guarantees" packets transfer in order. UDP on the other hand, does not give any guarantees to your packets.

Why is it "faster?" The header overhead does not sum up to be a huge amount, however, there's no need to wait for an ACK every packet for UDP. The cost of UDP is that each individual program needs to now keep track of packets received. Not only that, UDP packets can be duplicate, delivered out of order, or even dropped - so you have to take care of these cases. Plus the sender can silently disappear and the receiver needs to deal with this case too, versus the sender just being slow or not have much data. It's much more trouble than claiming it's "faster"....

I wasn't looking at all that. I was just looking at the header and the handshake stuff. Thanks for refreshing my memory! I tried writing a program for the PS3 once that used the broadcast address. By default though, the firewall on Ubuntu (which I don't run anymore, but did at the time), was blocking it and I could never figure out how to allow the traffic through! I figured if I was having this much trouble with it, then general users probably would as well. I was sending UDP packets to the broadcast address. The idea was the PS3 was the server and the Linux box was the client. I was trying to get it where the PS3 didn't need to know the IP address of the Linux box at first. The PS3 program would send a UDP packet on the broadcast address. The Linux box would look for this packet, once it found it, it'd send back a packet with the IP address of the Linux box, and then it'd establish a connection.

I'm sure it was just something stupid I was doing. I'm still not an expert at iptables, even though I've learned a lot over the years.

... IPV4 and IPV6 software necessarily needs to be written differently as your userland code now needs to handle 128 bit IP addresses versus the 32 bit addresses in IPV4. However the API, at least for Un*x, is still very similar. Writing code that listens to both IPV4 and IPV6 means you'll have some duplicate code that needs to be maintained separately.

Yeah. Maybe even quadruple code. One for IPv4 in Linux, one in Windows, one for IPv6 in Linux, one in Windows. With this Weller preheater I got, there's a serial port on the back. I contacted Weller and they gave me the protocol information and I can communicate with it with a custom program I've written. I wanted to release it open source but have a Windows version and a Linux version. I didn't want two separate programs though. And I didn't want to have a whole bunch if #IFDEFs for Windows and Linux. I found a third party C library that really helped. The library takes care of all the serial stuff so I don't have to. Maybe for my BBS program, I can just write a small library that does all the dirty work. That might be the way to go. Or, because the code for IPv4 and IPv6 is very similar, maybe I can just have a function for the IPv# specific stuff, and then where the code is the same, a different function or something. You know, like

int initiate_ipconn(int ip_ver) {
  if(ip_ver==4) {
    /* Do IPv4 specific stuff */
  } else {
    /* Do IPv6 specific stuff */
  }

  /* Do IP version independent stuff */

  return 0;
}

[Per Hansson]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

I’ll bet every single one of them are also still running V4 ... and they wont stop either ... see the problem here?

Now if you wanna tell me that pretty soon, no new networks will be able to get a V4 address block because ICANNA wont issue them anymore ... well then yeah ... they would at that point be forcing the change by being dicks about it. lol

It sounds to me like you have not read the news in the last few years.
IANA ran out of address space in 2011.
It does not mean the world is ending but just like what has been stated before the whole 32bit IPv4 address space (4.2B addresses) is not usable.
The actual number of usable addresses is something like 3.7B, but that also is a high number as you can't perfectly distribute out those addresses.

And it creates also another issue which is routing related: before you could look at an IP and know which country it belonged to.
Not anymore, the address space is so segmented that addresses very close in numbers can be on different continents.
This makes routing tables very big and complicated.

This is what IPv6 is here to solve, it's to make the address space utilization easier, not for you but for the routers that do the actual work.

https://en.wikipedia.org/wiki/IPv4_a...tes_and_impact

https://www.iana.org/assignments/ipv...ress-space.xml

[eccerr0r]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

- A class A is a /8, a class B is a /16, and a class C is /24. As seen in my case I have a /29 partial C so it is possible to break a network outside of an A, B, or C class. However every break, as said, will have more overhead addresses wasting more addresses, so it's a tradeoff (plus the breaks need to be an integer power of 2). To better save IPV4 addresses everyone needs to be on the same and huge subnet, and people know what messes comcast has to deal with.

Because of the inefficiencies and address space segregation necessary, there's no way to constantly dole out ipv4 addresses. V6 is unfortunately needed and I will not stand behind NAT even if I get hundreds of hack attempts every day.

- There is indeed more chatter with IPV6, but that's because it's autoconfigure plus the additional bytes for specifying the addresses - and it's only on the local segment. But still, the amount of chatter packets is not that large - it's mainly the router advertising itself to all on that segment. The endpoint machines do not need to keep chatter the router unless it has real data it needs to route out.

- UDP does not mean faster than TCP. Both are running over IP so there's always that limit, but the main difference is that TCP holds state and "guarantees" packets transfer in order. UDP on the other hand, does not give any guarantees to your packets.

Why is it "faster?" The header overhead does not sum up to be a huge amount, however, there's no need to wait for an ACK every packet for UDP. The cost of UDP is that each individual program needs to now keep track of packets received. Not only that, UDP packets can be duplicate, delivered out of order, or even dropped - so you have to take care of these cases. Plus the sender can silently disappear and the receiver needs to deal with this case too, versus the sender just being slow or not have much data. It's much more trouble than claiming it's "faster".

In fact UDP could be slower on most WAN except those extremely reliable networks. Most people can't even gauge this because TCP hides all the network garbage that occur, actually fairly frequently - people (and software writers) don't typically see any wrong order, most of the lost, or duplicate packets.

Plus there are those ISPs and NAT routers that automatically drop UDP routing information after a while because they can't tell when you're done - when FIN packets for TCP are loud and clear - just to make sure they don't get completely overwhelmed with holding too much state. You'll really learn to hate these routers. I suppose there's a reason why NFS was rewritten to allow TCP connections though historically it's been completely UDP.

- IPV4 and IPV6 software necessarily needs to be written differently as your userland code now needs to handle 128 bit IP addresses versus the 32 bit addresses in IPV4. However the API, at least for Un*x, is still very similar. Writing code that listens to both IPV4 and IPV6 means you'll have some duplicate code that needs to be maintained separately.

---

Yes I run my own home mail server. For now I'm stuck with using IPV4 mail because reverse V4 DNS works fine. I had to explicitly disable IPV6 because googooeverywhere insists on using ipv6 and not fall back to ipv4 when an address has both A and AAAA records. Maybe someday my ISP will fix reverse V6...

Godaddy was actually tunneling IPV6? That would have been neat, then again Hurricane still has ipv6 tunneling for the public? I happen to have 6RD tunneling with my ISP so that should work out better, except they don't know how to do V6 PTR records.

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

...I have a dual stack IPV4 and IPV6 setup through 6RD, but unfortunately there are problems with IPV6. There are a lot of people who don't have IPV6 setup properly, and I can't even set up my IPV6 properly because my ISP can't set up IPV6 reverse DNS properly. As far as I know I am fully functional listening to IPV6 requests but since reverse DNS does not work, people reject connecting to my machine... Figures, have to use IPV4 which reverse DNS does work.

And yeah, I virtually have a /64 or something like that IPV6 chunk... No way I can use that many addresses...

Isn't reverse DNS mainly used for mail servers and anti-spam techniques? Are you running some sort of mail server?

I had some issues setting up IPv6 on my server originally, but smart people at cPanel helped. Turned out I wasn't doing everything properly. However, I kinda cheated. I used cPanel (a kinda graphical sysadmin tool). I didn't set it up through the console, I set it up through cPanel.

GoDaddy originally had IPv6 all setup and in place, but they got rid of it, saying not enough users were using it! I thought that was the stupidest thing they could do!!!! For what it's worth, with GoDaddy, you cannot setup reverse DNS. It's because they have this system in place that's supposed to prevent people from sending massive amounts of spam. All outgoing mail goes through one of their relays. That's why when you do a reverse dns lookup on a GoDaddy server, you see something like ip-<ipaddress>.secureserver.net, instead hostname.domain.com

I want to say even if you rent a dedicated server from them, you don't have the ability to setup reverse DNS. That's one of the reasons I moved away from them. The other, being the IPv6 stuff.

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

Well then I guess thats gospel on the issue ... two of the most heavily used ISPs on the planet having address issues .. hmmm. ....

Its not those guys is the non-ISP related companies who have the majority of the addresses... SEARS and Roebuck for example ... I guarantee you they aren't using 1/10th of the addresses they have assigned to them ... or Walmart ... or pick any other company besides an ISP who has to actually re-assign IP addresses and will naturally be struggling because everyone else has all the damn addresses... My solution would give a hell of a lot more addresses back into the hands of the ISPs who actually need them.

Think about it ... the number of NON-Internet service related companies FAR outnumbers the internet service companies ... and yes, they can apply for and get IP addresses straight from the source just like the ISPs do ... there is a shit ton of waste going on trust me.

Even if companies had to give back the IPv4 addresses they weren't using, that's still a temporary solution. Eventually, we'll be back to where we are now, with running out of IPv4 addresses. It might take a good amount of time, but I think we just need to sit back down at the drawing board and either revise IPv6 or, like you suggested earlier, come up with a better protocol all together (I'm thinking something like Spork/IP, that sounds good!)

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

The US registrars ran out of IPv4 addresses almost two years ago.
Several other areas, notably in Asia ran out much earlier than that.
https://en.wikipedia.org/wiki/IPv4_address_exhaustion

IPv6 was launched 5 years ago.
https://en.wikipedia.org/wiki/World_...Pv6_Launch_Day

Worldwide the adoption currently sits at 15%
Notably in countries with big populations the adoption is much higher:
In the US and Germany the adoption is at 30% in India it's at 22% and in Japan at 17%
https://www.google.com/intl/en/ipv6/statistics.html

Finally the IPv6 space is 340 trillion, trillion, trillion addresses long.
So your math is off by a factor of say a few dozens orders of magnitude!

I think one of the problems with IPv4 was like in Deposit. The PCs had public IPs for whatever reason. We did business with banks and FedEx, but mostly big banks. Banks that small banks do business with (ie, Juniper). Class C was too small, Class B gave us waaayyy more IP addresses than we'd ever use. But I don't think there was away to get a partial class B network. I don't think they could have asked the US registrar's for 1,000 class B public IP addresses, for instance. I think you need to get the full 65,536 addresses. That's a lot of wasted IPv4 addresses!

[Spork Schivago]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

...so what I'm getting at is, wont the OS handle it for the code you're writing?

One of the long-term projects I'm working on is written in C (not C++). The code for the IPv4 stuff is different than the IPv6 stuff. Here's a link that can show you the differences in the code:

http://www.lugod.org/presentations/ipv6programming/

Thats a loooong eventually since we technically can give up probably half or more of what we have assigned right now...

Yes, we're a long ways away from using up all the IPv6 addresses. But I think the overkill was the whole point. Remember computers back in the day? First, 64K was more than enough memory! Then, Commadore, they came out with computer that had a whopping 128KB of RAM! When MS-DOS came out, I believe (don't quote me), but Microsoft said hey, let's make the OS support an insane amount of memory so we'll never need to worry about it again, let's make the OS support 640KB of RAM! But that wasn't enough as time went on.

Same with hard disks. Many times, developers would increase the maximum size hard disk that was supported.

With IP addresses, we're seeing more and more "smart" devices. Cellphones, light bulbs, water heaters, electrical receptacles, blu-ray players, televisions, video game consoles, watches, etc that can all connect to the internet. NAT works. I am with you with the IPv6 though on how it was setup. I think it needs a lot of improvements. I didn't know it was just thrown together in a panic, but that makes since. Ever connect to a network that is IPv6 only and fire up something like Wireshark in promiscuous mode? There's so much more "chatter" than there is with IPv4. I couldn't imagine what it'd be like in a large corporation! The congestion must be nuts!

...UDP is far better... but what we need is some genius to write the killer protocol that will speed everyone up by like 50% and win us over that way...

It would be nice if we could find a better protocol than TCP/IP. UDP is better for speed, but has it's downfalls as well. With the UDP, there's no error checking. The server sends the packets. If the recipient misses a packet, too bad. There's no way for the recipient to request the server to send that packet again. With TCP, the recipient creates a connection, does all the handshaking stuff, and then asks for the TCP packet. The server sends the TCP packet. The recipient responds and lets the server know it's received the packet. If it doesn't receive it, it'll ask the server to send again until it does.

This is just a simplified explanation and it's been quite a long time since I took the CCNA classes, so I might have some of it wrong. If anyone sees something wrong with my explanation, feel free to chime in and correct me!

...Any communication network that has no backup links is a network with a single point of failure...

This was my fault. I should have been a bit more clear. With all the residential setups I've seen, there's always one point of failure. What I meant was here, with Spectrum, we've lost television but still had internet and phone. According to Empire, if one service goes out, they all go out.

....The other benefits of fiber ... not susceptible to magnetic interference... lightening ... weather balloons ... northern lights whatever ... your data isn't going to be jumbled up over a few magnetic waves.

Even though fiber isn't susceptible to magnetic interference, having fiber for telephone doesn't necessarily mean you won't experience interference though. For example, I call my buddy Jason who has a cell phone. The cell phone is broke. Every 30 seconds or though, he cuts out. No matter where he is. Even when he has full bars. There's a lot of good benefits for fiber and I really want it. I honestly cannot see how it could be worse than Spectrum. I cannot see how any company could be much worse than Spectrum. I just really need the IPv6.

Sounds like a paid VPN service whos primary goal is to mask your ip address so you can do whatever without being traced ...

It's just a Virtual Private Server. I pay them money each month and they give me access to one of their virtual machines. I can do what I want with it. If it breaks, they fix it for free (if the hardware fails). It's like renting a VPS from GoDaddy, but I feel Linode is a much better company. I decided to install CentOS 7 on my VPS. It stores my website right now for that project I'm working on. Freetime is sparse now and I don't have much time to work on it though. Eventually, my server will act as a central server / database for my program. When someone decides to run my program, it'll connect to my server to communicate with other people running the same software.

My Linode server has an IPv4 address and an IPv6 address. Essentially, I'd want it where once the tunnel is setup, for testing purposes, I could, on my personal computer at my house, write software where it tries to connect via IPv4 and fails, but if I try to connect with IPv6, it'll work. Do you understand? I need to make sure when I'm testing the IPv6 stuff that we're not falling back on IPv4 and I just think we're connecting with IPv6. That's all.

Just had a thought ... have you considered using a virtual environment to do this in?...

Yes, I have thought of this. I actually have a system here that will be essentially a clone of the Linode server. It won't be a virtual server, it'll be a physical server. A development server. cPanel already said they'll give me a free developers account so I can install cPanel on this server for free and not have to pay for it. That way, when I test stuff for the website or for cPanel, instead of just putting it on the production server, I can test various things locally first. If it works well, then I'll move it to the production server.

I don't actually want to run a server from my house at this point in time where users will connect to. I want them connecting to my Linode server, no a server in my physical house.

Interesting story about the BBS code. Unfortunate he would not release it, I wonder why he was hanging onto it. How did that benefit him?

Personally, what I think happened, back then, when BBSes were popular, with certain ones, you could pay for the source code for a certain amount of money (maybe 50$ for the BBS, but if you wanted the source code, 100$). However, you were only allowed to modify the source code for your BBS. You weren't allowed to change it and then sell it. I suspect the person of this BBS, the one who won't release the source code, he paid money for the source code and just changed it to make it look like a different BBS and started selling it for profit. I think when it spread across the states to NY (he lived on the other side of the country), they got scared and shut it down, figuring they might get caught. I might be wrong though. I've read a good bit about this developer and when he wrote the BBS, he was young. A person said the first BBS they ran was this BBS software. Some high school kids in Washington (the state) wrote it and where selling it for 50$ a copy.
The BBS supports a scripting language, something kinda like BASIC. I find it hard (but not impossible) to believe that a high schooler could write an interrupter, especially one that supports GOTO statements. That's not the easiest thing to do. I've tried. There's a good bit to it. I've talked to people who could write this scripting language. Computer scientists with a PH.d, for example. They said they'd be willing to do it, but because of the work involved, something like this would cost me around 10,000$.

I was under the impression that even modern programming languages can do Branching such as the go to statement. For example, I believe in Java all you have to do is create a label, something like :myLabel then you simply use the goto statement two branch to that label.

What language do you primarily code in?...

I primarily code in C. And languages like C do support goto statements (although a lot of good programmers will look down upon programmers using goto statements). The problem isn't the programming language that I write the BBS in. The problem is the BBS itself has a built-in programming language. You customize the BBS by modifying these .MNU files. The .MNU files are a bit like a BASIC file. You write code in those .MNU files. Stuff like

if %userlevel%==9 goto sysop
if %userlevel%==8 goto cosysop

Using goto's in a programming language isn't hard. Writing a program that can interrupt a language is a lot harder. I've purchased some books to help with this. One is the infamous red dragon book (Compilers: Principles, Techniques, and Tools). This is a programming book that's designed to teach you how to write compilers, interrupters, etc. Stuff like Python would be an interrupted language. That means the source code doesn't get compiled. A program interrupts it. I sometimes refer to these as scripting languages. Essentially, I need to write my BBS software so it's a BBS but also something like Python or BASIC or whatever. I compile the code, but it loads text files that have a custom language inside and the BBS software can understand and does stuff differently based on whatever's inside that file.

Do you understand? I can PM you and send you more details and examples if you want. I know it's a bit of information to take in and I'm not doing the best job explaining all of it.

[eccerr0r]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

There are a lot of reserved addresses in ipv4 so the 4 billion number is not accurate:
- Network and broadcast addresses. Ultimately every time a network is broken down to a smaller chunk, more addresses are 'wasted' to segregate traffic from each piece. Normally a class C network subnet would only have 2 wasted addresses (broadcast and network address), but with my small /29 chunk, this same class C is shared among 32 subnets, which thus means 64 addresses are wasted. Plus people tend to not have machines also serving as routers, so more than 64 addresses are wasted - more like 96 out of 256. That's 32x more wasted IP addresses to make sure people play nice with each other and not do deliberate IP collisions.
I think the /29 is probably a fairly extreme case, but that's what it is. /30s are also possible which is even more wasteful.
- Reserved address spaces like private address spaces and multicast addresses. Granted private address space is the saving grace for a lot of IPV4 troubles, but it will only go so far. I chose to have static IP so my voice can be heard.

I have a dual stack IPV4 and IPV6 setup through 6RD, but unfortunately there are problems with IPV6. There are a lot of people who don't have IPV6 setup properly, and I can't even set up my IPV6 properly because my ISP can't set up IPV6 reverse DNS properly. As far as I know I am fully functional listening to IPV6 requests but since reverse DNS does not work, people reject connecting to my machine... Figures, have to use IPV4 which reverse DNS does work.

And yeah, I virtually have a /64 or something like that IPV6 chunk... No way I can use that many addresses...

[goontron]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

Why couldn’t they have just added a couple of octets to v4 and called it good? Why was it necessary to go from having this many addresses

[SIZE="5"][RIGHT]4,294,967,296
340,282,366,920,938,463,463,374,607,431,768,211,456[/RIGHT][/SIZE]

to THAT MANY ... I mean for the love of all that is good ... we don't even have a word for a number that big ... its 3.4 x 10 ^22

few people would even comprehend just how large of a number that is ... and it makes the addresses rather damn large and complicated ...

Why you so upset about my dislike of it anyways?

Not sure if you are responding to me or not with that last line.... But anyway, i'm a sysadmin for a local datacenter. We are having issues with IPv4 addresses. I would love if sears gave up some IPv4! We are looking into IPv6, but so little of the customer endpoints support it, it's pointless.

I would love to move to IPv6.... Get a chunk of a couple million addresses for pennies! You have documentation for the addresses yeah? We do for out IPv4 even, No problem! But the ISPs aren't pushing it out fast enough.

[EasyGoing1]

Re: Trippling internet speed with pare of AA batteries, ethernet cable and electric t

Why couldn’t they have just added a couple of octets to v4 and called it good? Why was it necessary to go from having this many addresses

[SIZE="5"][RIGHT]4,294,967,296
340,282,366,920,938,463,463,374,607,431,768,211,456[/RIGHT][/SIZE]

to THAT MANY ... I mean for the love of all that is good ... we don't even have a word for a number that big ... its 3.4 x 10 ^22

few people would even comprehend just how large of a number that is ... and it makes the addresses rather damn large and complicated ...

Why you so upset about my dislike of it anyways?