Lab Scenario 2

CertificationZone Page 1 of 3 http://www.certificationzone.com/studyguides ./?Issue=39&IssueDate=06-01-2001&CP= 11/06/01 Date of Issue: 06-01-2001 Lab Scenario 2 by Annlee Hines The Scenario Proposal 1 Cost of proposal 1 Proposal 2 Cost of proposal 2 Analysis Proposal 1 Proposal 2 Solution Note for the Financially Astute The Scenario Archimedes and Sons, Inc. is a developer and distributor of science and mathematics training programs aimed at community colleges and technical schools. Rather than provide CD-ROMs, which the technically-adept students tend to copy prolifically for their friends (thereby depriving A & S of their revenue), the firm provides interactive web-based training to user accounts (which are still abused, but less so). The high demand for A & S's training comes from their intensive use of animated, three-dimensional graphics which show the pieces of a solution coming together into a whole, and three-dimensional models of complex structures. These graphics are not downloaded to the student; the student sends commands (such a rotation or skew) to the server, which generates and sends the video images as they are created. Business has been growing rapidly and customers are beginning to complain about the speed of response to commands. A & S knows the problem is not in the servers; when the programs are run locally, they are as fast as ever. While nothing can be done about the students' end of a connection, A & S has decided to upgrade its WAN connections to the Internet. A & S currently has three T-1s feeding their ISP on a loadsharing connection (traffic is shared across all three T-1 lines; connections are not discretely tracked), running Frame Relay. Each T-1 costs $2650/month; they lease their edge router for $215/month. Global Networks, A & S's carrier, has responded with two proposals: Proposal 1 Proposal 1 is to increase A & S's bandwidth by doubling the number of T-1s and running a compression algorithm on the video before streaming it over Frame Relay. Their existing edge router does not have the capacity to support this higher load, so they must buy or lease another one, dividing the server load between the two routers, or use a larger and faster single router. The sales engineer has advised A & S that, if their business grows much larger, it will become cheaper to lease a T-3 than to keep adding T-1s. This would occur at the eight T-1 point. Cost of proposal 1 Six T-1s at $2650/month each, plus two edge routers at $215/month each. Instead of leasing the routers, they can purchase one larger router for $6335, including a software upgrade from their current release. If they grow to the point where they have to lease 8 T-1s ($21,200/month), it will become cheaper to lease a T-3 ($20,500/month). While they are not there yet, they are not that far away, either. Proposal 2 Proposal 2 requires adding only one more T-1, but they will change from Frame Relay access to ATM. Since A & S does not have a good estimate of how much of its outgoing traffic is video and how much is web traffic, the initial proposal is a three-month trial of two T-1s carrying rt-VBR traffic and two carrying UBR. Accounting analysis will show if this allocation needs adjusting. Cost of proposal 2 CertificationZone Page 2 of 3 http://www.certificationzone.com/studyguides ./?Issue=39&IssueDate=06-01-2001&CP= 11/06/01 The two rt-VBR T-1s will cost $3800/month each and the two carrying UBR will cost $3000/month each. The ATM edge switch will take care of the frame/cell conversion. It may be leased for $435/month or purchased outright for $8250. Which proposal should A & S accept, and why? Analysis A & S has two clearly separate categories of traffic: normal web traffic and interactive video. Expanding their Frame Relay bandwidth may ease congestion on A & S's end of the link for a time, but expansion (business is growing) is likely to lead to the problem recurring. A one-time problem would not be a serious blow to A & S's reputation, but a fix that failed in short order would leave them looking less than competent. Cost of the first proposal is another factor. Proposal 1 l Six T-1s @ $2650/month is $15,900/month, or $381,600 for two years l Leasing two edge routers @ $215/month is $10,320 for two years l Purchasing one large edge router is $6,335. If they scale up to the T-3, the cost will grow to $492,000 for data services, plus the cost of a larger router (lease at $385/month or buy for $7950). Best case total cost over two years is thus $391,920 if they lease a second edge router or $387,935 if they purchase the one larger router. Proposal 2 Proposal 2 offers to segregate the delay-sensitive video traffic from the delay-tolerant web traffic and apply QoS (by treating them as rt-VBR and UBR, respectively). By segregating the traffic, video will not be delayed by large web packets and will be more responsive. Less bandwidth will be needed to assure timely delivery of delay-sensitive traffic. Cost of proposal 2 is much more favorable: l Two rt-VBR T-1s @ $3,800/month + two UBR T-1s @ $3,000/month is $13,600/month, or $326,400 for two years. l Leasing the edge switch costs $435/month or $10,440 for two years. l Purchasing the edge switch costs $8,250. Total cost over two years is $336,840 if they lease the edge switch, or $334,650 if they purchase the switch. If, after the 3-month trial, they need more rt-VBR bandwidth, the third T-1 can carry some or all of its traffic as rt-VBR. The worst case price is all traffic as rt-VBR. The third T-1's price changes from ($3000/mo*24 mo) to ($3000/mo*3 mo + $3800/mo*21 mo), an increase of $16,800. That brings the two-year total to $353,640 (lease) or $351,450 (buy), which is still much more favorable than proposal 1. Solution The ATM proposal is more expensive in every individual component. But, because it is more efficient -- and thus requires fewer resources to do the job (and do it more reliably, thanks to QoS) -- it actually costs less, overall, to implement. Choose proposal 2. Note for the Financially Astute The CFO, never a patient man, scathingly corrects your analysis, reminding you that monies spent over time should not be compared in nominal dollars, but rather in dollars converted to their present value. He reruns the numbers and arrives at the same answer, proposal 2 (though like all accountants, he lists an expense as a negative value, pointing out the obvious matter that an expense is subtracted from earnings). CertificationZone Page 3 of 3 http://www.certificationzone.com/studyguides ./?Issue=39&IssueDate=06-01-2001&CP= 11/06/01 [IE-WANS-LS2-F03] [2001-06-01-01] Proposal 1 PV Option $15,900/mo + $430/mo for 24 months ($361,065.18) Lease $15,900/mo + $6335 now ($357,892.64) Buy Proposal 2 $13,600/mo + $435/mo for 24 months ($310,321.48) Lease $13,600/mo + $8250 now ($308,953.39) Buy Switch one T-1 from UBR to rt-VBR after 3 months ($331,361.36) Lease ($329,825.56) Buy Copyright © 2001 Genium Publishing Corporation . http://www.certificationzone.com/studyguides ./?Issue=39&IssueDate=06-01 -20 01&CP= 11/06/01 Date of Issue: 06-01 -20 01 Lab Scenario 2 by Annlee Hines The Scenario Proposal 1 Cost of proposal 1 Proposal 2 Cost of proposal 2. $15,900/mo + $6335 now ($357,8 92. 64) Buy Proposal 2 $13,600/mo + $435/mo for 24 months ($310, 321 .48) Lease $13,600/mo + $ 825 0 now ($308,953.39) Buy Switch