3 ms·
"DNS was invented so that you don't have to memorise IP addresses." Prior to DNS it was necessary to have all the IP addresses one needed stored in a local fil
by 1vuio0pswjnm7 3y ago
"DNS was invented so that you don't have to memorise IP addresses."
Prior to DNS it was necessary to have all the IP addresses one needed stored in a local file, HOSTS.TXT. It was not necessary to to memorise the addresses stored in this file. Memorisation of IP addresses was never a requirement to use the internet.
To this day, /etc/hosts allows me to avoid memorising IP addresses. No DNS required. But more importantly it allows me to avoid _having to type_ either IP addresses or domain names. I can assign short strings in place of typing addresses or domain names.
For example,
echo 127.0.0.1 localhost lo|sed -i -e 1r/dev/stdin -e 1N /etc/hosts
Now I can type "lo" without having to type "127.0.0.1" or "localhost"
echo 5.179.1.4 ftp.lanet.lv ftp1|sed -i -e 1r/dev/stdin -e 1N /etc/hosts
Now I can type "ftp1" instead of having to remember 5.179.1.4 or ftp.lanet.lv
echo ls |ftp ftp://ftp1/pub
There are some useful IPv4 addresses that I did memorise over the years.
For example, by memorising a root server address such as 198.41.0.4 I can lookup the IP address for any domain name, using non-recursive queries to only authoritative DNS servers, without using a DNS cache like 8.8.8.8 or some other third party DNS "service". All I need is a stub resolver, like dnsq, drill or dq.
By memorising the IP address of a .com nameserver, such as 192.5.6.30, I can lookup any .com domain name using the same non-recursive methodology. All the third party DNS caches could be poisoned, have outages or whatever. I would not need any of them to continue using the web.
And by memorising the address for getting a copy of the root zone, i.e., 192.0.32.9, I can avoid having to make DNS queries to find out the namservers for any TLD. I can use HTTP or FTP to retrieve the zone data.
ftp -4 ftp://192.0.32.9/pub/root.zone
As it happpens I store lots of historical DNS data so if there was a massive DNS outage I can still continue using the web uninterrupted. Day-to-day I'm not using remote DNS much. I'm using stored DNS data retrieved in bulk at some time point in the past. For the data I retrieve, very few addresses change often if at all.
I like IPv4. I think it will be around for the duration of my lifetime. It works everywhere and I like reliability. I disable IPv6 on the computers I can fully control. (Obviously "smartphones" are excluded from the list, that's why they suck as computers.) When compiling software that has "--disable-IPv6" as a configure option, I always disable IPv6. I also disable getaddrinfo() when it's an option. I prefer gethostbyname().
- orangeboats 3y agoThe last paragraph is confusing. Why still disable IPv6 when you are obviously not concerned by memorization of long addresses?
- 1vuio0pswjnm7 3y agoWhile I am not concerned with memorising every IP address I use, I do type and search IP addresses quite frequently; I work with addresses at the command prompt as parameters to command line utilties and in files, such as /etc/hosts and much larger files that contain hundreds of thousands of IP addresses, often using an editor such as ed, sed, vi or custom filters I have written in C. Needless to say, IPv4 addresses are easier to work with. When I check weather I use an IP address not a domain name. This can be quite useful on a smartphone where I generally cannot control DNS. This is an IPv4 address. As one might imagine, I have memorised it. Often this will work on public wifi without having to "sign in". I like IPv4.