Pages

Jumat, 20 November 2015

Analisa Jaringan IT

Mr. Bob is an IT director at Trust Oil and Gas, a medium-sized petrochemical company based in Houston. It also has operations in the Gulf of Mexico and in South America. Mr. Bob is in charge of the network infrastructure, including routers and switches. His group includes personnel who can install and configure Cisco routers and switches.
The Trust Oil and Gas CIO wants to begin migrating from the voice network to an Unified Communications (UC) solution to reduce circuit and management costs. Existing data WAN circuits have 50 percent utilization or less but spike up to 80 percent when sporadic FTP transfers occur.
Mr. Bob hands you the diagram shown in Figure 1. The exiting data network includes 35 sites with approximately 30 people at each site. The network is Multiprotocol Label Switching (MPLS) WAN, with approximately 200 people at the headquarters. The WAN links range from 384-kbps circuits to T1 speeds. Remote-site applications include statisti- cal files and graphical-site diagrams that are transferred using FTP from remote sites to the headquarters.



Mr. Bob wants a UC solution that manages the servers at headquarters but still provides redundancy or failover at the remote site. He mentions that he is concerned that the FTP traffic might impact the VoIP traffic. He wants to choose a site to implement a test before implementing UC at all sites.
The following questions/directives refer to this case study:
1.    What are the business requirements for Trust Oil and Gas?
2.    Are there any business-cost constraints?
3.    What are the network’s technical requirements?
4.    What are the network’s technical constraints?
5.    Approximately how many IP phones should the network support?
6.    What type of UC architecture should you propose?
7.    What quality of service (QoS) features would you propose for the WAN?
8.    What PoE recommendations would you make?
9.    Would you propose a prototype or a pilot?
10.    What solution do you suggest for voice redundancy at the remote sites?
11.    Diagram the proposed solution.

Answer:

1.    The company wants to provide services in a coverged network
2.    The solution should provide reduced costs over the existing separate voice and data networks.
3.    The technical requirements are as follows:
a.    Provide IP telephony over the data network.
b.    Provide voice redundancy or failover for the remote sites.
c.    Prevent FTP traffic from impacting the voice traffic.
4.    The technical constraint is as follows:
       Call-processing servers need to be located at headquarters.
5.    There are 200 IP phones at headquarters, and 35 * 30 = 1050 remote IP phones, for a total of  1250   IP phones.
6.    Propose the WAN centralized call-processing architecture with a CallManager (CM) cluster at headquarters.
7.    Use low-latency queuing (LLQ) on the WAN links to give the highest priority to voice traffic.  Then define traffic classes for regular traffic and FTP traffic. Make bandwidth reservations for the voice traffic and maximum bandwidth restrictions for the FTP traffic. Call Admission Control (CAC) is recommended to limit the number of calls from and to a remote site.
8.    All hubs need to be replaced with switches. All switches should be replaced with PoE switches to provide power to future IP phones and wireless access points. All new switch purchases should be PoE-capable LAN switches.
9.    To prove that calls can run over the WAN links, implement a pilot site. The pilot would test the design’s functionality over the WAN with or without FTP traffic.
10.    Recommend the use of Survivable Remote Site Telephony (SRST) to provide voice services  in the event of WAN failure, and reroute calls to the Public Switched Telephone Network (PSTN).
11.    This figure, which shows headquarters and two remote sites for clarity. This architecture is duplicated for all remote sites. Each site uses a voice router that is connected to both the IP WAN and the PSTN. SRST provides voice survivability in the case  of WAN failure. A CM cluster is implemented at the headquarters. The CM servers are in the data center in a redundant network.



0 komentar:

Posting Komentar