Computer Network midterm
T/F the transport layer sits on top of the network layer, and provides its services using the services provided to it by the network layer. Thus it is important that we know what is meant by the network layer's best effort delivery service
True. The transport layer's best effort doesnt provide much service
DNS functions. Match the function of a server to a given type of DNS server in the DNS server hierarchy.
Question 1. Authoritative DNS server: Provides authoritative hostname to IP mappings for organization's named hosts. 2. Local DNS server: Replies to DNS query by local host, by contacting other DNS servers to answer the query. 3. Top Level Domain (TLD) servers: Responsible for a domain (e.g., *.com, *.edu); knows how to contact authoritative name servers. 4. DNS root servers: Highest level of the DNS hierarchy, knows how to reach servers responsible for a given domain (e.g., *.com, *.edu).
Components of packet delay. Match the description of each component of packet delay to its name in the pull down list.
.0033 secs
The peer-to-peer (P2P) paradigm. Which of the characteristics below are associated with a P2P approach to structuring network applications (as opposed to a client-server approach)?
1. There is a server with a well known server IP address. 2. There is a server that is always on. 3. HTTP uses this application structure.
Computing Packet Transmission Delay (3). Consider the network shown in the figure below, with three links, each with the specified transmission rate and link length. Assume the length of a packet is 8000 bits.What is the transmission delay at link 2? link1 1: Transmission rate = 10Mbps; Link length = 2Km link 2: Tr = 100 Mbps; Ll= 1000Km link 3: Tr:100Mbps; Ll= 1Km
8 x 10^(-5) secs
T/F. When multiple UDP clients send UDP segments to the same destination port number at a receiving host, those segments (from different senders) will always be directed to the same socket at the receiving host.
True. UDP demultiplexing happens solely on the basis of destination port number. Thus, segments with the same dest. port num at the host will have their data demultiplexed to the same socket
Access network per-subscriber speeds. Match the access network with the approximate speeds that a subscriber might experience. (Note: if you look these up, do so in the 8E textbook, slides,or video -- not in the 7E or earlier versions, since link access speeds are always increasing over the years).
1)Ethernet - Wired. Up to 100's Gbps per link. 2)4G cellular LTE - Wireless. Up to 10's Mbps per device. 3)802.11 WiFi- Wireless. 10's to 100's of Mbps per device. 4)Cable access network- Wired. Up to 10's to 100's of Mbps downstream per user. 5)Digital Subscriber Line-Wired. Up to 10's of Mbps downstream per user.
Computing throughput: a simple scenario. What is the maximum throughput achievable between sender and receiver in the scenario shown below? sender-> link capacity 1.5Mbits/s ->router -> link capacity:10Mbits/s -> receiver
1.5 Mbps
When an application uses a UDP socket, what transport services are provided to the application by UDP? Check all that apply.
Best effort service. The service will make a best effort to deliver data to the destination but makes no guarantees that any particular segment of data will actually get there.
Link Transmission Characteristics. Which of the following physical layer technologies has the highest transmission rate and lowest bit error rate in practice?
Fiber optic cable
Streaming video definitions. Match the definition/function of an element or approach in a networked streaming video system, with its name.
Question 1. Chunk: A unit of video, each of which may be encoded at multiple different rates, stored in different files. 2. Manifest: A file containing the location and encoding rate of files corresponding to video segments in a video. 3. DASH: An approach that allows a client to adapt the encoding rate of retrieved video to network congestion conditions. 4. Enter deep: A CDN approach that stores content in access networks, close to clients.
The HTTP GET. What is the purpose of the HTTP GET message?
The HTTP GET request message is used by a web client to request a web server to send the requested object from the server to the client.
Conditional HTTP GET. What is the purpose of the conditional HTTP GET request message?
To allow a server to only send the requested object to the client if this object has changed since the server last sent this object to the client.
T/F. Is it possible for two UDP segments with source port 5732 to be sent by different processes at the sending host.
False. Each UDP socket on a local host must have a unique local port number associated with it.
The DNS authoritative name server. What is the role of an authoritative name server in the DNS? (Check all that apply)
It provides the definitive answer to the query with respect to a name in the authoritative name server's domain.
What is meant by transport-layer Multiplexing?
Taking data from one socket(one of possibly many sockets), encapsulating a data chunk with header information-thereby creating a transport layer segment- and eventually passing this segment to the network layer.
Where is transport-layer functionality primarily implemented?
Transport layer functions are implemented primarily at the hosts at the "edge" of the network
Why Web Caching? Which of the following are advantages of using a web cache? Sselect one or more answers.
1. Caching generally provides for a faster page load time at the client, if the web cache is in the client's institutional network, because the page is loaded from the nearby cache rather than from the distant server. 2. Caching uses less bandwidth coming into an institutional network where the client is located, if the cache is also located in that institutional network.
Packet switching versus circuit switching (1). Which of the characteristics below are associated with the technique of packet switching
1. Data may be queued before being transmitted due to other user's data that's also queueing for transmission. 2.This technique is used in the Internet. 3.Resources are used on demand, not reserved in advance. 4. Congestion loss and variable end-end delays are possible with this technique.
T/F The Transport layer provides for host-to-host delivery service.
False. The network layer provides for host-to host delivery service, while the transport layer provides for process-to-process, or application-to-application delivery.
CDNs. What approach is taken by a CDN to stream content to hundreds of thousands of simultaneous users?
Store/serve multiple copies of videos at multiple geographically distributed sites.
"HTTP is stateless." What do we mean when we say "HTTP is stateless"? In answering this question, assume that cookies are not used. Check all answers that apply.
An HTTP server does not remember anything about what happened during earlier steps in interacting with this HTTP client.
What's in the DNS type A resource record? What information does the type "A" resource record hold in the DNS database? Check all that apply.
1. A hostname and an IP address.
Why DNS Caching? What is the value of caching in the local DNS name server? Check all that apply.
1. DNS caching provides for faster replies, if the reply to the query is found in the cache. 2. DNS caching results in less load elsewhere in DNS, when the reply to a query is found in the local cache.
E-mail delays. How many RTTs are there from when a client first contacts an email server (by initiating a TCP session) to when the client can begin sending the email message itself - that is following all initial TCP or SMTP handshaking required?
3
What is an HTTP cookie used for
A cookies is a code used by a server, carried on a client's HTTP request, to access information the server had earlier stored about an earlier interaction with this person. [Think about the distinction between a browser and a person.]
HTTP/2 versus HTTP/1.1: object download delays. Consider a client and a server, separated by an RTT of 4 time units. The client makes a request for 4 objects at t=0. O1 consists of 10 frames, O2 and O4 each consist of 1 frame, and O3 consists of 2 frames. In the HTTP/2 example shown below, the server is transmitting frames to the client in the order O1, O2, O3, O4 (as long as there are frames of type i to transmit, and when not the server just moves on to a frame from object i+1 mod 4). Each frame takes 1 time unit to transmit. Under HTTP 1.1 (not shown below), the server would send O1, O2, O3, O4 in that first-come-first-served (FCFS) order, sending each object in its entirety before moving on to send the next object in that order.Let's define the object download delay as the time from when an object is requested (at t=0 below) to the time that object is received in its entirety. What is the average object download delay (the sum of the four object download delays divided by 4) under the HTTP/2 object frame transmission order shown below and under HTTP/1.1 O1, O2, O3, O4 object transmission order?
Average object download delay under HTTP/1.1: 16.0, under HTTP/2: 10.5
DNS in Action (1). Suppose that the local DNS server caches all information coming in from all root, TLD, and authoritative DNS servers for 20 time units. (Thus, for example, when a root server returns the name and address of a TLD server for .com, the cache remembers that this is the TLD server to use to resolve a .com name). Assume also that the local cache is initially empty, that iterative DNS queries are always used, that DNS requests are just for name-to-IP-address translation, that 1 time unit is needed for each server-to-server or host-to-server (one way) request/response, and that there is only one authoritative name server (each) for any .edu or .com domain. Consider the following DNS requests, made by the local host at the given times: t=0, the local host requests that the name gaia.cs.umass.edu be resolved to an IP address. t=1, the local host requests that the name icann.org be resolved to an IP address. t=5, the local host requests that the name cs.umd.edu be resolved to an IP address. (Hint: be careful!) t=10, the local host again requests that the name gaia.cs.umass.edu be resolved to an IP address. t=12, the local host requests that the name cs.mit.edu be resolved to an IP address. t=30, the local host again requests that the name gaia.cs.umass.edu be resolved to an IP address. (Hint: be careful!) Which of the requests require 8 time units to be resolved?
1. The request at t=0. 2. The request at t=1. 3. The request at t=30.
How many calls can be carried? Consider the circuit-switched network shown in the figure below, with four circuit switches A, B, C, and D. Suppose there are 20 circuits between A and B, 19 circuits between B and C, 15 circuits between C and D, and 16 circuits between D and A. What is the maximum number of connections that can be ongoing in the network at any one time?
70
Routing versus forwarding. Choose one the following two definitions that makes the correct distinction between routing versus forwarding.
Forwarding is the local action of moving arriving packets from router's input link to appropriate router output link, while routing is the global action of determining the source-destination paths taken by packets.
Components of packet delay. Match the description of each component of packet delay to its name in the pull down list
Question 1.Processing delay: Time needed to perform an integrity check, lookup packet information in a local table and move the packet from an input link to an output link in a router. 2.Queueing delay: Time spent waiting in packet buffers for link transmission. 3.Transmission delay: Time spent transmitting packets bits into the link. 4.Propagation delay: Time need for bits to physically propagate through the transmission medium from end one of a link to the other.
What is meant by transport-layer demultiplexing?
Receiving a transport-layer segment from the network layer, extracting the payload(data) and delivering the data to the correct socket
The local DNS server. Check all of the phrases below that state a true property of a local DNS server.
The local DNS server can decrease the name-to-IP-address resolution time experienced by a querying local host over the case when a DNS is resolved via querying into the DNS hierarchy.
The client-server paradigm. Which of the characteristics below are associated with a client-server approach to structuring network applications (as opposed to a P2P approach)?
There is a server that is always on. A process requests service from those it contacts and will provide service to processes that contact it. There is not a server that is always on. There is a server with a well known server IP address. HTTP uses this application structure.
Manifest file. What is the purpose of a manifest file in a streaming multimedia setting?
To let a client know where it can retrieve different video segments, encoded at different rates
T/F. Is it possible for two TCP segments with source port 80 to be sent by different processes at the sending host.
True
What is a protocol? Which of the following human scenarios involve a protocol (recall: "Protocols define the format, order of messages sent and received among network entities, and actions taken on message transmission, receipt")?
-Two people introducing themselves to each other. -One person asking, and getting, the time to/from another person. -A student raising her/his hand to ask a really insightful question, followed by the teaching acknowledging the student, listening carefully to the question, and responding with a clear, insightful answer. And then thanking the student for the question, since teachers love to get questions.
What is DASH? In DASH (Dynamic, Adaptive Streaming over HTTP), a server divides a video file into chunks that ... (pick best completion from below)
... are stored, each encoded at multiple rates (video quality). The client plays the video chunk-by-chunk, with each chunk requested at encoding rate that fits the available bandwidth at the time.
Computing Packet Transmission Delay (1). Suppose a packet is L = 1500 bytes long (one byte = 8 bits), and link transmits at R = 1 Gbps (i.e., a link can transmit bits 1,000,000,000 bits per second). What is the transmission delay for this packet?
.000012 secs
Computing Packet Transmission Delay (2). Suppose a packet is L = 1200 bytes long (one byte = 8 bits), and link transmits at R = 100 Mbps (i.e., a link can transmit bits 100,000,000 bits per second). What is the transmission delay for this packet?
.000096 secs
Trying out traceroute. Perform a traceroute from your computer (on whatever network you happen to be on) to gaia.cs.umass.edu. Use traceroute (on Mac terminal) or tracert (on Windows command line) or tracepath (on a Linux command line). Enter the missing part of the name of the router just before the host gaia.cs.umass.edu is reached: [A].cs.umass.eduNote: Routing may change, so the answer here may not be correct anymore. Also, if you are a Verizon user, there are known problems using traceroute with Verizon - if traceroute shows you two hops only to gaia.cs.umass.edu or any destination, skip this question.
nscs1bbs1
T/F. When multiple TCP clients send TCP segments with the same destination port number to a receiving host, those segments (from different senders) will always be directed to the same socket at the receiving host.
False, TCP demultiplexes on the basis of 4 values:source and destination port num, source and dest. IP address. Therefore if two segments come from diff. IP addresses or have diff source port nums, their payload will be demultiplexed to diff sockets, ieven if the dest. port nums are the same.
Computing utilization (3). Consider the scenario shown below, with four different servers connected to four different clients over four three-hop paths. The four pairs share a common middle hop with a transmission capacity of R = 300 Mbps. The four links from the servers to the shared link have a transmission capacity of RS = 50 Mbps. Each of the four links from the shared middle link to a client has a transmission capacity of RC = 90 Mbps. Assuming that the servers are all sending at their maximum rate possible, what are the link utilizations of the client links (with transmission capacity RC)? Enter your answer in a decimal form of 1.00 (if the utilization is 1) or 0.xx (if the utilization is less than 1, rounded to the closest xx).
0.56
Computing utilization (2). Consider the scenario shown below, with four different servers connected to four different clients over four three-hop paths. The four pairs share a common middle hop with a transmission capacity of R = 300 Mbps. The four links from the servers to the shared link have a transmission capacity of RS = 50 Mbps. Each of the four links from the shared middle link to a client has a transmission capacity of RC = 90 Mbps.Assuming that the servers are all sending at their maximum rate possible, what are the link utilizations of the shared link (with transmission capacity R)? Enter your answer in a decimal form of 1.00 (if the utilization is 1) or 0.xx (if the utilization is less than 1, rounded to the closest xx).
0.67
HTTP/2 versus HTTP/1.1. Which of the following are changes between HTTP 1.1 and HTTP/2? Note: select one or more answers.
1. HTTP/2 allows a large object to be broken down into smaller pieces, and the transmission of those pieces to be interleaved with transmission other smaller objects, thus preventing a large object from forcing many smaller objects to wait their turn for transmission. 2. HTTP/2 allows objects in a persistent connection to be sent in a client-specified priority order.
TCP service. When an application uses a TCP socket, what transport services are provided to the application by TCP? Check all that apply.
1. Loss-free data transfer. The service will reliably transfer all data to the receiver, recovering from packets dropped in the network due to router buffer overflow. 2. Flow Control. The provided service will ensure that the sender does not send so fast as to overflow receiver buffers. 3. Congestion control. The service will control senders so that the senders do not collectively send more data than links in the network can handle.
What is a network of networks? When we say that the Internet is a "network of networks," we mean? Check all that apply (hint: check two or more).
1. The Internet is the largest network ever built. 2.The Internet is made up of access networks at the edge, tier-1 networks at the core, and interconnected regional and content provider networks as well
Computing utlilization. Consider the scenario shown below, with four different servers connected to four different clients over four three-hop paths. The four pairs share a common middle hop with a transmission capacity of R = 300 Mbps. The four links from the servers to the shared link have a transmission capacity of RS = 50 Mbps. Each of the four links from the shared middle link to a client has a transmission capacity of RC = 90 Mbps.Assuming that the servers are all sending at their maximum rate possible, what are the link utilizations for the server links (with transmission capacity RS)? Enter your answer in a decimal form of 1.00 (if the utilization is 1) or 0.xx (if the utilization is less than 1, rounded to the closest xx).
1.00
Packet switching versus circuit switching (2). Which of the characteristics below are associated with the technique of circuit switching?
1.Reserves resources needed for a call from source to destination. 2.Frequency Division Multiplexing (FDM) and Time Division Multiplexing (TDM) are two approaches for implementing this technique. 3. This technique was the basis for the telephone call switching during the 20th century and into the beginning of this current century.
HTTP 1.1 GET (re-request) with an embedded object. Suppose an HTTP client makes a first GET request to the gaia.cs.umass.edu web server for a base page that it has never before requested, which contains an embedded object, which causes the client to make a second GET request. A very short time later, the client then make a third GET request - for the same base page, with that third GET request having an If-Modified-Since field (as does the 4th GET request that the client makes for the embedded object). Neither the base object nor the jpeg object have changed. How many round trip times (RTTs) are needed from when the client first makes the third GET request (i.e., when it requests the base object for the second time) to when the base page and the jpeg file are displayed a second time, assuming that: any time needed by the server to transmit the base file, or the jpeg file into the server's link is (each) equal to 1/2 RTT the time needed to transmit an HTTP GET into the client's link is zero? the time needed by the server to transmit a reply that does not contain the base page or an embedded jpeg object is zero You should assume that persistent HTTP 1.1 is being used and should think about the consequences of the If-Modified-Since field being used in the third and fourth GET requests). You should take into account any TCP setup time required before an HTTP GET is actually sent by the client, the time needed for the server to transmit a requested object (see assumptions above), and any propagation delays not accounted for in these amounts of time.
2 RTT
Simple HTTP GET request response time. Suppose an HTTP client makes a request to the gaia.cs.umass.edu web server. The client has never before requested a given base object, nor has it communicated recently with the gaia.cs.umass.edu server. You can assume, however, that the client host knows the IP address of gaia.cs.umass.edu. How many round trip times (RTTs) are needed from when the client first makes the request to when the base page is completely downloaded, assuming the time needed by the server to transmit the base file into the server's link is equal to 1/2 RTT and that the time needed to transmit the HTTP GET into the client's link is zero? (You should take into account any TCP setup time required before the HTTP GET is actually sent by the client, the time needed for the server to transmit the requested object, and any propagation delays not accounted for in these amounts of time.)
2.5 RTT
Download delays for 100 objects (HTTP 1.1 with local web caching). Consider an HTTP 1.1 client and server. The RTT delay between the client and server is 2 seconds. Suppose the time a server needs to transmit an object into its outgoing link is 3 seconds. There is also a local web cache, as shown in the figure below, with negligible (zero) propagation delay and object transmission time. The client makes 100 requests one after the other, waiting for a reply before sending the next request. All requests first go to the cache (which also has a 2.0 sec. RTT delay to the server but zero RTT to the client). How much time elapses between the client transmitting the first request, and the receipt of the last requested object, assuming no use of the IF-MODIFIED-SINCE header line anywhere, and assuming that 50% of the objects requested are "hits" (found) in the local cache?
252 secs
Download delays for 100 objects (HTTP 1.1 with browser caching). Consider an HTTP 1.1 client and server. The RTT delay between the client and server is 2 seconds. Suppose the time a server needs to transmit an object into its outgoing link is 3 seconds, as shown below for the first of these 100 requests. You can assume that any other HTTP message not containing an object sent by the client and server has a negligible (zero) transmission time. Suppose the client makes 100 requests, one after the other, waiting for a reply to a request before sending the next request.Using HTTP 1.1, how much time elapses between the client transmitting the first request, and the receipt of the last requested object, assuming the client uses the IF-MODIFIED-SINCE header line, and 50% of the objects requested have not changed since the client downloaded them (before these 100 downloads are performed)?
352 secs
HTTP 1.0 GET with an embedded object. Suppose an HTTP client makes a request to the gaia.cs.umass.edu web server. The client has never before requested a given base object, nor has it communicated recently with the gaia.cs.umass.edu server. You can assume, however, that the client host knows the IP address of gaia.cs.umass.edu. Suppose also that after downloading the base file, the browser encounters a jpeg object in the base html file that is stored on gaia.cs.umass.edu, and therefore makes another GET request to gaia.cs.umass.edu for that referenced jpeg object. How many round trip times (RTTs) are needed from when the client first makes the request to when the base page and the jpeg file are completely downloaded, assuming the time needed by the server to transmit the base file, or the jpeg file into the server's link is (each) equal to 1/2 RTT and that the time needed to transmit the HTTP GET into the client's link is zero? You should assume that HTTP 1.0 is being used. (You should take into account any TCP setup time required before an HTTP GET is actually sent by the client, the time needed for the server to transmit the requested object, and any propagation delays not accounted for in these amounts of time.)
5 RTT
1.4-7 Computing throughput. Consider the scenario shown below, with four different servers connected to four different clients over four three-hop paths. The four pairs share a common middle hop with a transmission capacity of R = 300 Mbps. The four links from the servers to the shared link have a transmission capacity of RS = 50 Mbps. Each of the four links from the shared middle link to a client has a transmission capacity of RC = 90 Mbps.What is the maximum achievable end-end throughput (an integer value, in Mbps) for each of four client-to-server pairs, assuming that the middle link is fairly shared (divides its transmission rate equally) and all servers are trying to send at their maximum rate?
50
Download delays for 100 objects (HTTP 1.1). Consider an HTTP 1.1 client and server. The RTT delay between the client and server is 2 seconds. Suppose the time a server needs to transmit an object into its outgoing link is 3 seconds, as shown below for the first of these 100 requests. You can assume that any other HTTP message not containing an object sent by the client and server has a negligible (zero) transmission time. Suppose the client makes 100 requests, one after the other, waiting for a reply to a request before sending the next request.Using HTTP 1.1, how much time elapses between the client transmitting the first request, and the receipt of the last requested object?
502 secs
Download delays for 100 objects (HTTP 1.0). Consider an HTTP 1.0 client and server. The RTT delay between the client and server is 2 seconds. Suppose the time a server needs to transmit an object into its outgoing link is 3 seconds, as shown below for the first of these 100 requests. You can assume that any other HTTP message not containing an object sent by the client and server has a negligible (zero) transmission time. Suppose the client makes 100 requests, one after the other, waiting for a reply to a request before sending the next request.Using HTTP 1.0, how much time elapses between the client transmitting the first request, and the receipt of the last requested object?
700 secs
Packet switching or Circuit-switching? Consider a scenario in which 5 users are being multiplexed over a channel of 10 Mbps. Under the various scenarios below, match the scenario to whether circuit switching or packet switching is better.
Question 1. Neither work well in this overload scenario: Each user generates traffic at an average rate of 1.2 Mbps, but generates traffic at rate of 15 Mbps when transmitting 2. Circuit switching: Each user generates traffic at an average rate of 1 Mbps, generating traffic at rate of 1 Mbps when transmitting 3.Packet switching: Each user generates traffic at an average rate of 15 Mbps, generating traffic at rate of 1 Mbps when transmitting
True or False: On the sending side, the UDP sender will take each application-layer chunk of data written into a UDP socket and send it in a distinct UDP datagram. And then on the receiving side, UDP will deliver a segment's payload into the appropriate socket, preserving the application-defined message boundary.
True.