Jump to content
Sign in to follow this  
Guest Larry S

Question for all you computer experts

Recommended Posts

Guest christianholmes

Ok, I don't mean to sound condescending to anyone (I'm not pointing out anyone in particular), but I notice whenever somebody asks a technical question, several people jump in who all say they are correct (myself included). I just participated in another thread where this happened. So I want to know- how much do you really know? I've devised a question here that really can't be answered with an internet search- you have to know what you are talking about to answer it correctly (it's a devious one). Please keep in mind, I was a MS instructor, so I don't pretend to know everything, but I do know quite a bit about OS's and networks. Also keep in mind any REAL Cisco or MCSE could answer this- it's pretty standard stuff- if they're not too rusty.The winner gets an Nividia GeForce 4600Ti if they want it (you pay shipping- probably a few bucks). You have to explain your answer.Here it is:You have two LAN segments, connected by a router. You have to configure another machine on either side of the router, and you have to fill in the blanks. You've been giving a valid IP by the Enterpise Admin. You can't change the router settings, and there are thousands of machines on both LAN segments. Using the IP addresses, you have to figure out what subnet mask, and gateway to use for both new machines. They need to be able to communicate with all computers on their LAN segment, as well as the other side of the router. And yes, there is only one correct answer :)EDIT: Just to clarify, I've added the IP of one more machine on each segment.1)202.56.92.176 subnet= gateway=||------- 202.56.83.112|----| | router----||------- 202.56.48.252|1)202.56.56.156 subnet= gateway=As a bonus- tell me what the broadcast address is for both machines, and I'll pay shipping.Sometime tomorrow or Friday, if no-one gets it, I will explain the answer, for anybody who cares- you'll probably find it quite interesting.

Share this post


Link to post

Im no computer expert but the question is interesting. Anyways Im going to give it a shot.IP = 202.56.92.176subnet = 255.255.0.0gateway = 202.56.1.1thenIP = 202.56.56.156subnet = 255.255.0.0gateway = 202.56.1.1The broadcast address for each client its is own IP address.Oh well. It was fun trying to guess.Ed

Share this post


Link to post
Guest Matt Johnson

I'd argue that there's not nearly enough information here, and that there is more than one valid solution given the information we have!Multiple "conventions" exist for deriving gateway addresses. One is the second address in the CIDR block (immediately after the network address), the other is the last-but-one (so an address ending in .254 where CIDR 80.0/20Gateway: 202.56.64.1 OR 202.56.80.1Broadcast: 202.56.95.255"Southern" subnetNetwork: 202.56.32.0/19 OR 202.56.48.0/20Gateway: 202.56.32.1 OR 202.56.48.1Broadcast: 202.56.63.255/19 looks more likely if we really only have a two segment network though (having a gap in the network would be a really weird thing to do). Both are equally "valid" however and would encompass both IPs as given on both networks -- a /21 on either network would miss one IP in either of the subnets.Best regards,Matt

Share this post


Link to post

Ha! easy one. You'll have to ask tougher questions than that. I've got the answer but you're not going to like it :-)42 :-lolCheers,RogerI'm in Aus so donate the video card to the charity of your choice :-)

Share this post


Link to post
Guest metamarty

Haven't looked at the other answers, but here's my try:Because we have a router between the networks, there should be different IP networks on both sides, so we need at least different network addresses. Because there are thousands of machines on both end, the host part of the addresses should be at least 10 bits large but preferable larger. The first bit that differs between the given addresses is the 2nd bit in the 3rd digit of the IP addresses. BUT: using only up to the 2nd bit as the network address leaves the south end of the router with a zero subnet, and unless this is configured, it is not allowed, so I'll add 1 more bit to the netmask. In this case it would be 11111111.11111111.11100000.00000000 or 255.255.224.0.Using this netmask, the subnets would be:north: 202.56.64.0south: 202.56.32.0As a general convention, the gateway is the first host on the network so these should be:north: 202.56.64.1south: 202.56.32.1broadcast addresses are then:north: 202.56.95.255south: 202.56.63.255Hope this is correct. I'm a bit rusty in the networks department. It's been some years since I blasted hand crafted UDP packets on the school network to crash servers :-outta

Share this post


Link to post
Guest Larry S

ChristianI just gotta comment on this. No answer, just a comment.I did tech support for that company and OS that you mentioned for five years. My credentials may not be as impressive as yours are, but I'm way beyond needing any of the "for dummies" books. (and I do mean your credentials are more impressive than mine, nothing but true respect was intended there)I used to input on technical questions. If I wasn't sure my answer was correct, it was full of things like "I think" or "I may be corrected on this".I answered a hardware question a long time ago. My answer was backed up by two tech manuals from that company you mentioned and the company that made the hardware in question. The information I gave may not have been the only correct answer, but it was a correct answer. I was told by one of the "experts" that anybody who thinks that doesn't have a clue.Now I can take being told I'm wrong when I'm wrong, and if I'm given the correct answer, I appreciate it and try to learn from my mistakes. But there's a point where I draw the line.I read the posts, and learn from the people who know what they're talking about, and ignore the "experts". But I no longer contribute in the technical posts. You sound a little frustrated. Hope you have lot more patience than I do.Larry

Share this post


Link to post
Guest christianholmes

Hi everybody,Well firstly outstanding job on this. Record speed I might add. I kept the question open quite a bit longer than I originally thought, since none of the answers were exactly what MS looking for. Yes, although the IPs have changed this in an actual question that has since been discontinued. In fact they only asked this one on first few versions of the Server 2003 MCSE upgrade exams.Larry, I'm not really sure what happened with you, but I'm not really frustrated. I didn't do this as a real "contest" either- I just wanted to give incentive to have people participate. I thought it might be fun to start some REAL technical discussions- that is why I started this. I'm tired of the same old stuff, and I thought people might want to actually learn something interesting. I know that whenever I see an informative post about something, I like to read it and learn. I mean, how many posts do we see about video cards, and pixel shaders 2.0 vs. Godzilla?Matt- Good work- You are correct that there are technically a few ways to do this. However, there is only 1 correct answer based on "Microsoft reasoning"- you can't exclude information I didn't give you (what?)- basically when I said there are other computers, you don't know what the IPs are, so you have to cover EVERY POSSIBLE IP address that will still allow you to solve the problem. Believe me, they really think like that. I'm curious why you didn't chose a /18? Why not use the largest available network segment? Either way, you know your stuff! And, yes a router can be anything- but even recently Cisco has started using the first available IPs.Samaritano- Nice try, but that won't work. I'll explain in a bit-Marty, good job also, and correct- but not correct in MS land :(. I would love to work with the both of you in a second though!Before I explain, let me say that teaching these classes was TOUGH- Imagine a bunch of egotistical network admins sitting in a room when you walk in. Government contractors, young hot-shots, and even one guy that ran IT for a HUGE American Bank (he had 3000 people working under him). The only way to get these people to listen to you, was to shut them up from day 1. So, I would take a question like this- walk in, and write it on the board. I'd then say I had to finish something and leave for about 20 minutes. Naturally, I'd sit in an adjacent room, and listen- trying to pick out the ones that I knew were going to be difficult. Well, at first you'd hear the typical excuses "been awhile", "I only know Cisco", blah blah. After the twenty minutes, I'd go back in to find that only a handful had the correct answer, and the ones who didn't wouldn't give me a problem the rest of the course. Most of the time the questions were easier than this. Maybe a DNS problem, or visual basic problem. Believe me, when I went through the course I was humbled very quickly. Listening is much harder than speaking.OK- so here is a little lesson in subnetting for anyone who wants it- you'll probably have to read this to understand the answer.First, forget what you know about Classful Subnetting. Historically, classful subnetting meant this:Class A subnet = 255.0.0.0 or 'slash' 8 - /8Class B subnet = 255.255.0.0 or /16Class C subnet = 255.255.255.0 or /24A class A subnet is any IP that the first octet is in the range 1 to 127A class B subnet is any IP that the first octect is 128 to 191Class C is 192 to 223Class D is, by some conventions used for multicast. Multicast is basically a protocol used for video conferencing and other high-bandwidth applications.This is fine for smaller networks, but it is actually inefficient. You may use a class C network at home (255.255.255.0 or /24) but do you really need 252 valid IP addresses?using CIDR, or classless subnetting, you can grab only what you need. You could, for example, use a subnet of 255.255.255.252, which would give you an IP for two computers. If one was the gateway, that's all you need.Now, if you don't really understand how /16 relates to 255.255.0.0- crack open a cold one- I should explain binary math and how a subnet actually affects routing.You know what a byte is- at least you've heard the term.A byte is 8 bits. A bit is a logical operator that is either on or off. It is a base 2 number. In computer world- a bit is represented by a 1 or a 0.so a byte looks like this: 00110101notice each position can only be 0 or 1. Now, you know how some many things in computers deal in multiples of 256?Guess how many combinations of 1s and 0s there are in a byte? Yep- 256 possible combinations (0-255)So how do we convert a binary byte into decimal? THIS IS REALLY EASY!Each position in the byte represents a value. Start at the right hand side. That position has a value of 1. Now, each number to left of that is worth twice as much. so, 1,2,4 etc.Each position in a byte is worth this:Position: 8 7 6 5 4 3 2 1Value: 128 64 32 16 8 4 2 1Now, just add the value for each position that is a 1 and not zero. (1 = on, 0 = off Ever see the computer power switches with a 1 and 0?)So, our example was:00110101That means:32+16+4+1= 53Here are some more: 10001000= 128+8= 13611111110= 128+64+32+16+8+4+2= 254Now, this is how you need to think of subnet masks:A Class C, or /24 or 255.255.255.0 subnet mask is this:11111111.11111111.11111111.00000000twenty four ones. Note that in a subnet, all 1s and 0s are consecutive- you can't use 10011101.00101001.01110011.00000000- you will get an error.There is a great "cheat sheet" for subnetting:128 192 224 240 248 252 254 255128 64 32 16 8 4 2 1 the numbers on top represent the decimal equivilent of the rightmost non-zero octect in a subnet, while the bottom represents the number of host for that network.This means that if we have 11111111.11100000.00000000.00000000 that the chart relates to the second octect- since it's not all zeros. If you have a classful subnet, like 11111111.00000000.00000000.00000000 then the chart represents the first- but you shouldn't need the chart here- all 1's = 255 and all 0's = 0.so if we use 11100000, then we count over 3 from the left. our subnet would end in .224 A .224 network would support 32 hosts (assuming this is the fourth octet. BTW- Each byte is an octet in case you weren't clear on what I meant)a subnet of 11111111.111111111.11111000.00000000 (/21) looks like this:255.255- count over 5 for the third octet- .248.0Now the cool part is that like the above example, you can tell exactly how many hosts are available for any given subnet by using the chart.If our subnet mask is 255.255.255.248 - then just go to the chart, find 248, and look at the number below it.248 = 8 hosts per network.255.255.255.128128 = 128 hosts per network. Think about it- 128 is 1/2 of 256, so each network is exactly half.so 255.255.248.0 = 8 x 256 gives us 2048 addresses (not all useable however). Note that we multiply by 256 since the 3rd octet changed. If it were the fourth octet, then it would be 8 hosts.Memorize this chart- you will amaze people with your speedy math skills :)Now let's take an IP address of 192.168.1.1 and 192.168.1.193 and show how the subnet mask compares them11000000.10101000.00000001.00000001 = 192.168.1.111000000.10101000.00000001.11000000 = 192.168.1.193So if we put our /24 mask against this:11000000.10101000.00000001.xxxxxxxx11000000.10101000.00000001.xxxxxxxxdo these match? Yes- so therefore they are on the same subnet.Now, let's add a few bits to our subnet mask, and make it a /26. This means we have 26 1s and only 6 0s, or:11111111.11111111.11111111.11000001Now, let's compare our IPs against this:11000000.10101000.00000001.00xxxxxx11000000.10101000.00000001.11xxxxxxDo they match? No!So we have a 26 bit subnet mask- How many hosts does that allow us?look at the chart- 64. So our valid network IDs would be.0, .64, .128, and .192Why? let's look at those 2 bits in the fourth octect: .11xxxxxxxnow translate the binary maths to decimals:for the numbers 0-63, the first two bits will be .00xxxxxxxxfor the numbers 64-127, this will be .01xxxxxxxfor the numbers 128-191, this will be .10xxxxxxxxand for 192 and above, it is .11xxxxxxxxso you can see that any two computers with IPs 0-63 will have IPs that match based on this subnet mask, and they can see each other. The same for the other three ranges.So, in this case, we have 4 seperate networks. I hope you are following along at least slightly- this is much easier in person :)So, here is how the question is answered:We know immediately we can't use classful subnetting:255.0.0.0, 255.255.0.0, and 255.255.255.0 won't allow for the proper communication.Using 255.0.0.0 won't allow the machines to talk to the network on the other side of the router (a router is a default gateway- traffic will only be sent to it when the IP you are trying to get to is not on your local subnet). This would tell the internal routing that all these IPs are local to each other, and the traffic wouldn't be routed.255.255.0.0 won't work for the same reason.255.255.255.0 won't work for another reason- it is too small! using this would mean the two machines that are local to each wouldn't be able to communicate properly, as the subnet mask would send the packets to the router. You don't want the packets going to the router for machines on the same LAN segment, so you need a subnet mask that will satisfy both criteria.Each LAN segment should use all available IPs, and each segment should be consecutive. And although not required, you should use the first available network ID.Looking at our IP addresses, and trying different subnet masks, we want to see what mask will allow the machines on the same segment to talk to each other, but route when necessary.using a /18, or 255.255.192.0, we can have 64 x 256 hosts per network. That means our network IDs are 0, 64, 128, and 192.In this case, /18 will include both 92.176 and 83.112 since the valid IPs are 64.1 through 127.255The other network would then be 0.1 through 63.255 which includes 48.252 and 56.156Also, you use the first available network address this way, which is good for setting up consistency for the future.1)202.56.92.176subnet= gateway=||------- 202.56.83.112|----| | router----||------- 202.56.48.252|1)202.56.56.156subnet=gateway=Subnet=255.255.192.0 or /18gateway1=202.56.64.1networkID1=202.56.64.0subnet broadcast1=202.56.127.255gateway2=202.56.0.1networkID2=202.56.0.0subnet broadcast2=202.56.63.255OK, now does anybody else have a good question?And does anybody want the card? Matt?

Share this post


Link to post
Guest

42 is indeed correct. It's the ultimate answer to the ultimate question :-lol

Share this post


Link to post
Guest Matt Johnson

Nah, keep the card. :)/18 is certainly possible, but not all of us are endowed with huge IP blocks. :-)--M

Share this post


Link to post
Guest metamarty

But what is the Microsoft way? The only difference I see is that you allow a subnet zero on the south side. Cisco reasons that this creates confusion since the subnet would have an identical address to the original subnetted network address. That's why the command "ip subnet zero" should be used to allow zero subnets on cisco routers. I'm not familiar with Microsoft but I took a CCNA exam 3 years ago.

Share this post


Link to post
Guest

I think he means "think like a typical candidate for the MCSE exam would" which is "apply the rules without thinking about the consequences".

Share this post


Link to post
Guest metamarty

That might well be the case, but that's why I would like to know Microsoft's rules regarding networking. I know Cisco's exams really test your knowledge. They try to bring you down by inventing commands that do not exist. The natural response for people that only learn from braindumps is to guess what the command would do, but in this case you have to rely on your knowledge and say that it doesn't exist.I have another question but this one is for programmers. It's not that hard, but I feel this question is very important. The question is:Why would you have a hard time writing a LALR parser for Lisp?I found out the hard way because when I started creating MartyLisp as a hobby project of mine, my first approach was to create the parser.I feel this question is very important to modern day programmers as it provides insight in the knowledge that people have of the building blocks of moders software systems. In my working field I encounter a huge amount of people (most of them have all kinds of certifications) who miss a lot of knowledge about fundamental computer science. Most of them claim that this is all "history" or not relevant anymore, but if you thing that Lisp is dead, think again. The HTML your browser parses right now and the XML you use IS lisp, but only written in a different syntax.Understanding the Lisp family languages also helps you understand the different approaches to working with sets of structured data, so I feel it's vital to broaden your knowledge in this field.I have nothing to give away but respect for the question, but please give it a try! Here's the question again:Why would you have a hard time writing a LALR parser for Lisp?

Share this post


Link to post
Guest christianholmes

Honestly- The Microsoft thinking is, quite, frankly- "You have a huge network". Seriously- If there is one thing that they consistently preach, it is that people routinely have 30,000 hosts on their networks. Why not? All those copies of XP must be somewhere... Routing is very different between devices- but this question is more theoretical. I love CISCO stuff, and it performs MUCH more consistently that MS stuff, but in theory this should work on both. I wasn't going to put a brain dump question on here, as it would about .1 second for somebody to find the answer.I'm not "a really excited user" of MS or anything, but I think they do some things pretty well.Not VPNs though :) I haven't been able to connect to my RRAS server for about a week now. I know if I reboot it, it will work- but that's just annoying. I just setup an endpoint on my PIX, but where the hell do I get the Cisco VPN client? I tried to get it off the website, but that's acting screwey- I log in, try to download it, and it says I need to log in. Oh well, I'll just have SBC send me one since I bought the stuff from them. It's kind of urgent since I have 10 remote sales people who need it, but they just started reporting the problem a few days ago (I thought it was just me- shows how much work they've doing :) )Unfortunately the MS certifications are available to anyone who wants to memorize a few brain dumps. There are some 400,000 MCSEs out there, but I don't know too many who could tell you how to configure a DNS server to automatically return HOST records to an intranet server on a client's local subnet (It was a test question, and a setting that most people have never seen). And the prep material is sub-par. You might find one short sentence in an MS book that explained it. With the Server 2003 exams, they really increased the pool of questions, and changed the tests quite often so you couldn't just screw your way through them. I know for a fact that in the first 6 months the test were out there were less than 5,000 people certified on Server 2003. In fact, you recieved a special certification for being one of the first 5000 (I think it was 5000- maybe less) and the offer stood for 7 or 8 months at least. Do you think they care though? They make $75 for every test that VUE gives. 7 tests for each person. That's quite a killing. The thing is just a scam to get people to sell their products for them. No wonder they give MCSE's the entire MSDN subscription for $300- it's just so you can get familiar with it.As for the Lisp question- I've never used Lisp. The "oldest" language I used was SAS. I thought Lisp was basically a parsing language? Parses are highly recursive, so maybe that'a limitation in Lisp- I don't know.

Share this post


Link to post
Guest

I severely doubt whether Microsoft created Windows XP with its use in large routers as an intended mode of operation...Similarly Cisco doesn't use the custom OS for their routers in desktop workstations...

Share this post


Link to post
Guest christianholmes

I'm not sure what you're talking about. Their assumption of large networks is why they like to use the largest subnets possible. It very common in MS land to have several Active Directory Forests spread across dozens of sites with multiple domain controllers in each site. Also web clusters- for intranets- is not uncommon in what they teach. XP as a router- no- but Server 2000/2003- You betcha! Who needs Cisco when you can do all your routing with MS! They support RIP v.2 and OSPF... so why use anything else? And with ISA server, you have your firewall, router, e-mail server, file and print server (with DFS and printer pooling), web server, certificate server, domain controller, DNS, WINS, and DHCP server all on one box! All you need is a quad P4 with 16 gigs of RAM and and a 2 terrabyte RAID 10, and you can get rid of those pesky (and small and non-impressive looking) 2600's and 515's.

Share this post


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  
  • Tom Allensworth,
    Founder of AVSIM Online


  • Flight Simulation's Premier Resource!

    AVSIM is a free service to the flight simulation community. AVSIM is staffed completely by volunteers and all funds donated to AVSIM go directly back to supporting the community. Your donation here helps to pay our bandwidth costs, emergency funding, and other general costs that crop up from time to time. Thank you for your support!

    Click here for more information and to see all donations year to date.
×
×
  • Create New...