Internet?

[KeepItSimpleStupid]

Disk drives are wierd. There is capacity and formatted capacity and marketing.
1000 units per K is better for marketing
1024 units per K isn't.

Memory is sold as 1024 units/K as far as I know.

Disk drives have raw capacity.
They have formatted capacity.
They have hidden bad block tables.
They have calibraion data.
They have partitions.

We need a number to buy a drive. That's going to be raw capacity.
We need a number to tell us what will fit on a used drive. What used to be called "Free blocks" or 512 block sectors way back when. Then things started to get wierd.
1 character files took up Sixteen - 512K blocks etc.

It's not an exact science. I once did a presentation (A few hundred people) on a the directory structure that was used on a small minicomputer. Back then I understood it. The OS was limited in storage and the largest integer. Growth in capacity was not built in. The same thing happened with DOS/Windows.

 

[Electroenthusiast]

It's not an exact science. I once did a presentation (A few hundred people) on a the directory structure that was used on a small minicomputer. Back then I understood it. The OS was limited in storage and the largest integer. Growth in capacity was not built in. The same thing happened with DOS/Windows.

Somewhat interesting topic to research...

 

[Sceadwian]

Kiss, don't even mention modern solid state disk drives where there's so much going on inbetween the actual storage element and the system controller itself as to blur the line between processing and storage. Nowdays everything has a processor more complicated than the ultimate state of the art machines did 10 years ago.

I think of the simplest form I can right now which is a flash memorty card, aside from the actual flash memory every SD card on earth has both a standard SPI interface, a high speed proprietary SD interface and virtually anything larger than the micro format also contains a USB slave controller for a controller-less PC interface

 

[Electroenthusiast]

Hi again, when you speedtest.net your internet speed, you see three different stats showing up. Download Speed, Upload speed and Ping. Whats this ping? How is it important?

 

[edeca]

Hi again, when you speedtest.net your internet speed, you see three different stats showing up. Download Speed, Upload speed and Ping. Whats this ping? How is it important?

It depends what they are showing as there are a number of ways of measuring it. It represents the "latency" or time delay between packets leaving the test machine and arriving at yours.

It is important because you could have a very high bandwidth link that is actually quite "slow". Consider a 1 gigabyte per second link between here and the moon. It might be capable of high data rates but it could still take a few minutes for data to arrive at the other end.

An analogy would be a very wide, very long water pipe. Lots of data can go IN one end every second, lots of data can come OUT the other end every second. But the total time for data to traverse from IN to OUT is not instantaneous.

 

[Sceadwian]

Just a minor technicality but ping time is not one way trip measure, it's the time it takes to get to a destination and back, it's called a ping like a sonar echo location ping.

Bad ping times can affect some protocols that rely on a large number of ack/nack requests for transfer verification, as it spends more time waiting for the ACK/NACK's that it does actually transferring data. It will affect general web browsing if it get up past around 100ms or so, 200ms or more can drastically affect how fast your connection 'feels' regardless of how high the transfer rate is, but you shouldn't see pings that high unless you use satellite Internet because the up-stream for satellite is a phone line.

Ping time is hyper critical for some online games like first person shooters and most real time multiplayer games.

 

[KeepItSimpleStupid]

Using the cmd prompt and using "ping <router address> -t -n 8192 "

CNTRL+C - stops this test

"ping 127.0.0.1" is also known as "localhost". It can tell you if your PC's TCP/IP stack is working.

Some, because of abuse, disable ping from outside of the network. My ISP disables windows traceroute, but if I go though some hoops and use a different packet, it will work. The ISP doesn't want you to know your public IP address. www.whatismyip.com will easily tell it to you. It changes at some rate depending on your ISP. If you rest your modem you will likely get another.

If you have large latencies, it can help you diagnose wiring issues. I still haven't figured out whether your public IP address and your Private IP address access the same wire segment or whether it's different if you access from outside the connection. I used to be able to do this, but I don't have the account anymore. I could log onto a Unix system and then ping back to me.

I just don't understand the router concept especially in my case. The router's config address is 192.168.0.1, but the router's address is 10.0.1.1 and the <public address> is something different.

 

[Sceadwian]

My ISP disables windows traceroute, but if I go though some hoops and use a different packet, it will work. The ISP doesn't want you to know your public IP address. www.whatismyip.com will easily tell it to you. It changes at some rate depending on your ISP. If you rest your modem you will likely get another.

Really? I've not found this to be the case myself. Your local IP address is always your public IP address around here at least on the cable or DSL services, although with DHCP leasing it can change (I don't think mine has since I got it) I guess it depends on the ISP.

It sounds like your ISP does some really strange routing, you don't technically have Internet access, you have access to a NAT service that does, where as the services around here provide true internet access, meaning the IP I have is a TRUE IP that routes directly to my cable modem. I'm NAT'd from there though as I have a house router, but with UPNP routers nowdays and software that knows how to use it you'll never even notice.

 

[KeepItSimpleStupid]

See, tracert disabled

Tracing route to www.l.google.com [74.125.131.103]
over a maximum of 30 hops:

1 * * * Request timed out.
2 * * * Request timed out.
3 * * * Request timed out.
4 ^C
C:\Users\Owner>tracert

Yep, I have access to a NAT service that does, the address the the ISP assigns is public.

See: https://joesbitbucket.blogspot.com/2006/10/linux-traceroute-vs-windows-tracert.html about tracert varients.

When I had Cygwin installed, I could get that version of tracert to work with checked options. It is annoying.

I should have said that my DSL modem has the config address of 192.168.1.1, BUT my router has the address of 10.0.1.1. The router does not have an HTTP interface. The DSL modem does. It's obvios that the DSL MODEM config address (http) is not on my network. The DSL modem is in BRIDGE Mode. I don;t understand how the mode actually works.

In order to config the modem, I have to connect via an Ethernet cable and put my laptop on the 192.168.0.x network.

I'm not sure if i could change the config address of the DSL modem to say 10.0.1.127 and be able to access it.

I'm being different:
I have a DSL modem operating in BRIDGE mode
I have a router capable of dial-up and DSL wireless internet. Wireless is broken and I haven't pulled the card. It's card based and 802.11b.
I have an Access point providing wireless services.
and I have a b/g Repeater.

Wierd things happen. One of them is if I start to connect near the repeater and then start walking with the laptop so it's in range of the base, it won;t connect automatically.

Very, very rarely wireless networking will totally quit and a reboot is the only fix (WIN 7). It used to happen on an XP machine and a Net stop Wireless zero Configuration and re-starting would always fix it. I can't find a fix in win 7. Fortunately it's very very rare. 2-3 times a year.

My version of Windows suffered from the bug that you could have a red X in the wireless ICON in the system try, but be connected wireless to the Internet. I fixed that with a group policy. Wireless connectivity from wakeup is typically slow, but it will connect unless I walk with the laptop during the connection process.



I know I can access devices from outside the home, or used to be able to, because I had my Slingbox and router configured properly with port forwarding.

Just not sure what the router does with a 192.168.1.x packet generated on the 10.0.1.x side. It doesn't appear to pass it through the router port. It is supposed to be non-routeable, but what device enforces that?

Not sure how and if adding static routes to Windows will make this work.

If I ping 192.168.0.0 and arp -a, it nothing is in the arp table

 

[Electroenthusiast]

@kiss, sceadwian, edeca
there arises two ques,
from what i understand, ping is the time required to make a 1 to-and-fro cycle. So does it depend on the distance between the server and my PC?
Since, all the signals travel in speeds of light, the highest possible ping is 50ms . But, i've heard a numbers more than that.

sceadwian, i don't understand this " ack/nack's ", neither google search helped me. Maybe, i need to search for a book titled like "Internet Demystified".

 

[KeepItSimpleStupid]

Routers process signals along the way, sometimes if the router cannot keep up the packet is discarded. It's not supposed to happen but it does.

Congestion at the router creates major delays
Crosstalk (cables to close together) or bad cables mess with signal integrity.

So, there are talks of HOPS which is each time a router says "It's not mine" and mearly forwards the packet. That's where "time to live" comes into play. It's the number of hops that you will allow to reach the destination.

Internet (TCP/IP) routes may not follow the exact same path from point A to point B and may not arrive at the same time. (UDP) is a bit different and would be used for streaming.

 

[edeca]

from what i understand, ping is the time required to make a 1 to-and-fro cycle. So does it depend on the distance between the server and my PC?
Since, all the signals travel in speeds of light, the highest possible ping is 50ms . But, i've heard a numbers more than that.

Well not all signals will travel that fast. Like KeepItSimpleStupid said there are propagation delays along the way. These are both electrical (e.g. signal in, signal out) but also as packets are processed.

However what is more important is the transmission medium. If your packets end up routed across a satellite link you can expect well over 100ms each direction (up to the satellite then back to earth). It doesn't matter if you don't have a satellite link yourself, if there is one anywhere in the whole path this will affect your round trip time (RTT).

Fortunately most intercontinental data will go via. undersea cables, see https://www.cablemap.info/ for a useful distraction.

 

[edeca]

Internet (TCP/IP) routes may not follow the exact same path from point A to point B and may not arrive at the same time. (UDP) is a bit different and would be used for streaming.

I have to slightly disagree here.. both TCP and UDP are carried over IP. In both cases individual packets may not take the same path and may arrive, may not arrive or may arrive out of order.

The difference comes at the high layer protocol, TCP is resilient to this at the protocol level, UDP is not and requires further steps to be taken in the application.

 

[KeepItSimpleStupid]

@edeca

Here is a discussion: https://www.skullbox.net/tcpudp.php

Overview
TCP (Transmission Control Protocol) is the most commonly used protocol on the Internet. The reason for this is because TCP offers error correction. When the TCP protocol is used there is a "guaranteed delivery." This is due largely in part to a method called "flow control." Flow control determines when data needs to be re-sent, and stops the flow of data until previous packets are successfully transferred. This works because if a packet of data is sent, a collision may occur. When this happens, the client re-requests the packet from the server until the whole packet is complete and is identical to its original.

UDP (User Datagram Protocol) is anther commonly used protocol on the Internet. However, UDP is never used to send important data such as webpages, database information, etc; UDP is commonly used for streaming audio and video. Streaming media such as Windows Media audio files (.WMA) , Real Player (.RM), and others use UDP because it offers speed! The reason UDP is faster than TCP is because there is no form of flow control or error correction. The data sent over the Internet is affected by collisions, and errors will be present. Remember that UDP is only concerned with speed. This is the main reason why streaming media is not high quality.

The above quote does not address the ordering of the packets, but if I were streaming data I might expect them to be in order. Note that in the above link a UDP packet does not have a sequence number.

How can you order something (UDP packet) without a sequence number?

 

[edeca]

The above quote does not address the ordering of the packets, but if I were streaming data I might expect them to be in order. Note that in the above link a UDP packet does not have a sequence number.

I should have been clearer, sorry. Packets are never guaranteed to arrive in order, that is what TCP can help to fix.

If an individual packet is fragmented (chopped into smaller chunks due to path MTU) then the IP layer will deal with it.

Regardless, both TCP and UDP sit atop IP.

How can you order something (UDP packet) without a sequence number?

Your application keeps track of packets and ordering using data from the payload of the UDP packet. This is exactly how streaming video and other protocols work. They use UDP due to the very low protocol overhead. A UDP header is fixed to 8 bytes, TCP headers are a minimum of 20 bytes and may be longer if there are options.

The difference with streaming is that lost packets are irrelevant very quickly. You don't want to waste CPU and bandwidth trying to recover data from a video frame which is now 5 seconds out of date!

The page you linked is misleading in a number of ways. Flow control is for congestion and is not inherently concerned with guaranteed delivery. Also I have no idea what the "collisions" mentioned are supposed to be. Packets either arrive or they don't (due to routing issues, congestion, etc.), this isn't particle science!

 

[RichTheDude]

As Edeca stated UDP can be reasonably reliable. Forward error correction can utilised to recover some errors, and extra redundancy can also be introduced.

In the 7 layer OSI model, the application layer can do whatever it fancies using UDP to make sure it gets the data it wants or is expecting.

 

[KeepItSimpleStupid]

Collisions defined: https://www.cisco.com/en/US/products/hw/modules/ps2033/products_tech_note09186a008009446d.shtml

 

[edeca]

Collisions defined: https://www.cisco.com/en/US/products/hw/modules/ps2033/products_tech_note09186a008009446d.shtml

Which is about Ethernet, thus totally irrelevant for almost all "internet" backbone routing! Ethernet is carrier sense, multiple access. Many of the OSI layer 2 protocols used between routers (such as ATM) are time division.

Like I said, the skullbox.net article mixes up a few terms and invents a few others.

My favourite book that covers much of this is by Richard Stevens: https://www.amazon.co.uk/TCP-IP-Illustrated-Protocols-The/dp/0201633469

 

[Sceadwian]

So does it depend on the distance between the server and my PC?
Since, all the signals travel in speeds of light, the highest possible ping is 50ms . But, i've heard a numbers more than that.

It has very little to do with the physical distance (down to the theoretical limit) it has to do with what and how many points on the internet it's routed through and what their traffic is like at the time of the connection attempt. Sure a transatlantic link is going to slow things down, but more from total traffic through that pipe than from any limitation on the signal propagation speed.

Keep in mind the physical distance between two points is nto the same as the distance the signal travels because the Internet is dynamically routed, it NEVER follows a straight line.

 

[Electroenthusiast]

Acc to above, the speed gets reduced, or there is a bit-rate constraint because the data is being routed. So, decrease(for an individual PC) in bit-rate happens. So, on what does ping depend? If both ping and bit-rate depend on the same aspect(due to routing), then why two different entities - bit-rate and ping?

In other words, consider a person practicing/competing in hurdles. Each hurdle obstructs his speed. At the end, only his speed gets reduced. Similarly, in Internet, only bit-rate gets reduced. I got the concept of 'ping', but didn't understand how the same thing affects both 'ping' and 'bitrate'. Hope you understood my point.