Transcript Switching
Network design Topic 1 Technical goals Agenda • Technical requirements – – – – – – – – – Scalability Availability Reliability Performance Security Manageability Usability Adaptability Affordability Scalability • How much growth a network design can support and continue to function properly • Can more users, applications, sites, WAN links and new technology be added without requiring a network redesign? • How much expansion is expected in the next 2 years? 5 years? • Issues that impact scalability – flat network topology of layer 2 switches • size of broadcast frames – selecting scalable routing protocols Availability • The time the network is operational and available to users • Hours operational/hours required to be operational – Measured in % uptime such as a goal of 99.999 % – If 8 hour * 5 day service is required per week • 8* 5= 40 hours per week • If network down for 30 mins within the required hours availability is (40-0.5)/40=99.75 % uptime – Redundancy is a means of providing availability • Duplicate devices and links to avoid downtime – Depends on reliability and resilience of components – Depends on disaster recovery and continuity Reliability • The accuracy, error rates, stability and the time between failures of equipment and technology • Mean time between failure (MTBF) – How long a component will last before failing – The mean time between service outages for the network • Mean time to repair (MTTR) – Availability = MTBF/(MTBF+MTTR) • For each critical application, how much money is lost from downtime per minute? Performance • Response time – Time between a request for a network service and a response arriving for the network service – Capacity or bandwidth • capability of the link and the switch to transfer data – Utilisation • the percent of total available capacity in use – Throughput • quantity of error free data successfully transferred – Accuracy • amount of useful traffic correctly transmitted – Efficiency • how much effort is required to produce a certain amount of data throughput, how much utilisation of network, CPUs – Delay or latency • time between transmission and delivery Security • Must protect from threats inside the network and outside the network • Should not disrupt normal business operations • Provide assurance for confidentiality and integrity – Identify assets and vital business functions – Analyse security risks – Develop security requirements Manageability • Does the customer have preferences? • Look for simple and centralised tools with consistent experience – Performance management • Traffic and application behaviour – Fault management • Detecting and correcting problems, trends – Configuration management • Remote operation of devices – Security management • Monitoring and testing protection methods – Accounting management • Accounting of network usage to allocate costs Usability • How easy it is for network users to access the network and services – Making network users job easier – User friendly and easy-to-use methods • • • • For mobile users as well From any location Automating and dynamic services and protocols Single logon – Strict security policies can inhibit usability • Tradeoffs may be justified Adaptability • Ability to adapt to change and to new technologies in the future – New protocols – New business practises – New standards and regulations • Adapt to changing environmental constraints – Changing traffic patterns – Changing priority requirements Affordability • How cost effective is the design? What is the customer’s budget? – To be cost effective the design should carry the maximum traffic for a given cost • Consider non-recurring equipment costs • Consider recurring network operations costs – Campus networks – low cost is more important – Enterprise networks – availability is more important – WANS – reoccurring cost of WAN circuits is a big expense • Minimise WAN traffic – routing protocols and management traffic – read only domain controllers • Improve efficiency with compression (WAN acceleration) • Use technologies that support oversubscription • Use less expensive DSL on public networks and VPNs Trade-offs • To meet high expectations for availability and performance – High cost circuits and redundant components required • To enforce strict security policies – Expensive monitoring and high tech equipment • To avoid network problems – Qualified and adequate staffing and retraining required • Tradeoffs must be made between cost and technical goals – Rephrase business goals as technical goals – Prioritise technical goals with customer Agenda • Technical requirements – – – – – – – – – Scalability Availability Reliability Performance Security Manageability Usability Adaptability Affordability