3. [10+10 20 pts] Consider the network configuration for a small college campus shown here. A web proxy server with a cache has been installed on the campus network and all the browsers are configured to send HTTP/10 requests to the proxy server. If the requested object is found in the proxy cache it is immediately retumed to the browser as an HTTP response. If the requested object is not found in the proxy's cache, the proxy server makes an HTTP/10request to the appropriate origin server and receives the object. It is then added to the proxy cache (if the object is cacheable) and the object is returned to the browser as an HTTP response. The browser programs do not cache HTTP responses and the proxy does not cache DNS resolution replies across web-pagc requests (however, it does store in memory DNS replies received while processing all embedded objects within the same web-page). The browsers and the proxy do not use parallel, persistent or pipelined connections Mcasurements of the HTTP requcstresponse characteristics have produced the following statistics: The mean (average) time to satisfy a browser's serverS public Internet HTTP10request with a response from the proxy cache is 10 milliscconds The mean time to satisfy a proxy's HTTP/ 1.0 request with a response from an origin server is 200 milliseconds router 2 .The web pages viewed by users contain an average o 6 objects (the base HTML file plus 5 embedded objects) average two different servers are contacted in router 1 network order to fetch all components of a page The campus on the campus network and is used by all resolvers on the campus network. Measurements of the DNS query/reply DNS server characteristics have produced the following statistics: local authoritative DNS server is also installed proxy server When the DNS server's cache is turned OFF, queries for typeA resource records (RR) need an average of 3 itcrated query/reply exchanges with non-local DNS servers and incur a mean delay of 600 milliscconds. When the DNS server's cache is tumed ON, the following statistics were measured for queries for type = A resource records (RR) * 70% were satisfied from the DNS cache with negligible delay 20% were satisfied by a single query/reply exchange with the DNS server authoritative for the name 5% were satisfied by two query/reply exchanges with a mean delay of 200 milliseconds 5% were satisfied by more than two query/reply exchanges with a mean delay of600 milliseconds. o o and incur a mean delay of 100 milliseconds o Cache Hit Ratios 0.8 06 The plot given here shows the expected relationship betwcen cache size and hit ratio for a proxy cache 04 0.3 02 10 12 In answering the following two questions, show how you derived your resul including all intermediate values used in your calculations. (a) With the cache tumed OFF in the local DNS server, approximately what size cache is needed in the proxy server in order for the mean time for users to fetch a complete web page (including all embedded objects) to be 600 milliscconds? (b) With the cache tumed ON in the local DNS server, approximately what size cache is needed in the proxy scrver in order for the mean time for users to fetch a complete web page (including all embedded objects) to be 600 milliseconds? 3. [10+10 20 pts] Consider the network configuration for a small college campus shown here. A web proxy server with a cache has been installed on the campus network and all the browsers are configured to send HTTP/10 requests to the proxy server. If the requested object is found in the proxy cache it is immediately retumed to the browser as an HTTP response. If the requested object is not found in the proxy's cache, the proxy server makes an HTTP/10request to the appropriate origin server and receives the object. It is then added to the proxy cache (if the object is cacheable) and the object is returned to the browser as an HTTP response. The browser programs do not cache HTTP responses and the proxy does not cache DNS resolution replies across web-pagc requests (however, it does store in memory DNS replies received while processing all embedded objects within the same web-page). The browsers and the proxy do not use parallel, persistent or pipelined connections Mcasurements of the HTTP requcstresponse characteristics have produced the following statistics: The mean (average) time to satisfy a browser's serverS public Internet HTTP10request with a response from the proxy cache is 10 milliscconds The mean time to satisfy a proxy's HTTP/ 1.0 request with a response from an origin server is 200 milliseconds router 2 .The web pages viewed by users contain an average o 6 objects (the base HTML file plus 5 embedded objects) average two different servers are contacted in router 1 network order to fetch all components of a page The campus on the campus network and is used by all resolvers on the campus network. Measurements of the DNS query/reply DNS server characteristics have produced the following statistics: local authoritative DNS server is also installed proxy server When the DNS server's cache is turned OFF, queries for typeA resource records (RR) need an average of 3 itcrated query/reply exchanges with non-local DNS servers and incur a mean delay of 600 milliscconds. When the DNS server's cache is tumed ON, the following statistics were measured for queries for type = A resource records (RR) * 70% were satisfied from the DNS cache with negligible delay 20% were satisfied by a single query/reply exchange with the DNS server authoritative for the name 5% were satisfied by two query/reply exchanges with a mean delay of 200 milliseconds 5% were satisfied by more than two query/reply exchanges with a mean delay of600 milliseconds. o o and incur a mean delay of 100 milliseconds o Cache Hit Ratios 0.8 06 The plot given here shows the expected relationship betwcen cache size and hit ratio for a proxy cache 04 0.3 02 10 12 In answering the following two questions, show how you derived your resul including all intermediate values used in your calculations. (a) With the cache tumed OFF in the local DNS server, approximately what size cache is needed in the proxy server in order for the mean time for users to fetch a complete web page (including all embedded objects) to be 600 milliscconds? (b) With the cache tumed ON in the local DNS server, approximately what size cache is needed in the proxy scrver in order for the mean time for users to fetch a complete web page (including all embedded objects) to be 600 milliseconds