Designing IP Addressing and Selecting Routing Protocols Part 3 — Transcript

Full transcript

1. 0:04designing a routing protocol deployment
2. 0:09this section first describes why certain
3. 0:12protocols are suitable
4. 0:14for specific modules in the enterprise
5. 0:16architecture
6. 0:18after that a discussion of the following
7. 0:20advanced
8. 0:21routing features like redistribution
9. 0:24filtering and summarization
10. 0:26will be covered
11. 0:29routing protocols in the enterprise
12. 0:31architecture
13. 0:33okay so the key point here is that
14. 0:36eigrp and ospf both
15. 0:39adapt quickly to changes and have
16. 0:42short convergence times therefore
17. 0:45they are suitable for use in the campus
18. 0:48core
19. 0:49okay we just have to remember that eigrp
20. 0:53is running only on cisco devices so if
21. 0:56you are using hybrid or different
22. 0:58vendors okay or different brands
23. 1:01of routers or network devices on your
24. 1:04network
25. 1:05then ospf is a good choice
26. 1:08okay so building
27. 1:12access within the enterprise campus
28. 1:14distribution
29. 1:15core so that is where we apply eigrp
30. 1:19or ospf now the decision of whether to
31. 1:22use
32. 1:23eigrp or ospf should be based on the
33. 1:26underlying physical topology
34. 1:28ip addressing equipment used
35. 1:32and possible issues related to routing
36. 1:34protocol in particular situation
37. 1:37okay now this figure here just
38. 1:40illustrates
39. 1:40routing protocols in the enterprise
40. 1:42architecture
41. 1:43including those recommended for campus
42. 1:46core
43. 1:48all right so internet connectivity
44. 1:51we can use static routing okay or bgp
45. 1:55now when to use bgp if you are
46. 1:57multi-homed
47. 1:59meaning the organization is connected to
48. 2:01different
49. 2:03isps or more than one isps okay
50. 2:06then it is recommended to use bgp
51. 2:11if your organization is single home to a
52. 2:13single isp
53. 2:15then we can deploy and use static
54. 2:18okay so within the enterprise
55. 2:22you can use also eigrp or ospf
56. 2:26same thing with the side to side vpn
57. 2:29one and man okay enterprise branch
58. 2:32within the branch okay you can use eigrp
59. 2:37or ospf there okay for connectivity
60. 2:40well you can use vpn
61. 2:44[Music]
62. 2:45router distribution so the following
63. 2:49are the possible reasons why you might
64. 2:51need
65. 2:52multiple routing protocols running at
66. 2:54the same time within your network
67. 2:56so basically the use of different
68. 2:59routing protocols
69. 3:00in a single autonomous or within the
70. 3:03organization it's not recommended
71. 3:05now that happens sometimes
72. 3:08when two organizations merged and this
73. 3:12two organization uses different routing
74. 3:14protocols
75. 3:15then that's the time we have to use
76. 3:17router distribution
77. 3:19okay so political issues
78. 3:23are included that's why there are some
79. 3:26organizations running different routing
80. 3:28protocols
81. 3:29or maybe they are using different uh
82. 3:32vendors
83. 3:33okay now for communications within the
84. 3:37cisco devices they are using eigrp
85. 3:40for communications within non-cisco
86. 3:42devices they are using ospf
87. 3:45and there is a need for this
88. 3:48group of networks to redistribute
89. 3:52or to connect then that is where router
90. 3:54distribution comes into play
91. 3:56okay so you are migrating
92. 4:00from an older igp okay to a new igp
93. 4:04multiple redistribution boundaries might
94. 4:06exist
95. 4:07until the new protocol has displaced the
96. 4:10old protocol completely
97. 4:12that's one of the reason okay so running
98. 4:15multiple routing protocols during the
99. 4:16migration is effectively the same
100. 4:19as network that has multiple routing
101. 4:20protocols running
102. 4:22as part of its design so you want to use
103. 4:26another protocol but have to keep the
104. 4:28old routing protocols because
105. 4:30maybe of the host systems needs
106. 4:34for example host based
107. 4:37or unix host based systems or routers
108. 4:40might run
109. 4:40only on rip then you don't have a choice
110. 4:44but to redistribute okay
111. 4:48some departments might not want to
112. 4:50upgrade
113. 4:51their routers to support a new routing
114. 4:53protocol
115. 4:55okay so if you have a mixed vendor
116. 4:58environment as i have mentioned earlier
117. 5:00you can use the cisco proprietary eigrp
118. 5:03routing protocol in the cisco portion
119. 5:06of the network and use the common
120. 5:08standard based routing protocol such as
121. 5:10ospf
122. 5:11to communicate with nancy devices
123. 5:15so that is when we use or consider
124. 5:18the utilization of route redistribution
125. 5:23now let's talk about the router
126. 5:25distribution direction
127. 5:26sorry distribution is often applied
128. 5:29between the campus core
129. 5:31and enterprise edge protocols now as
130. 5:34shown here in the figure
131. 5:35the distribution is possible in two ways
132. 5:38okay so you've got a one-way router
133. 5:42distribution
134. 5:43routing information is redistributed
135. 5:46from one routing protocol or domain to
136. 5:49another
137. 5:50but not vice versa so static or default
138. 5:53routes are required
139. 5:55in the opposite direction to provide
140. 5:57connectivity
141. 5:58okay now for the two-way router
142. 6:01distribution
143. 6:02routing information is distributed from
144. 6:05one routing protocol or domain to
145. 6:07another
146. 6:08and vice versa so static or default
147. 6:10routes are not required
148. 6:12because all routing information is
149. 6:14passed between two entities
150. 6:20how about router distribution in the
151. 6:22enterprise network
152. 6:24so redistribution is needed
153. 6:28in the building distribution layer when
154. 6:30different routing protocols or domains
155. 6:32exist in the building access
156. 6:34layer and the compass core so the
157. 6:37distribution
158. 6:38might also be needed when the campus
159. 6:40core
160. 6:41and enterprise edge including to and
161. 6:44from
162. 6:45one module routers from static or bgp
163. 6:48routes in the internet connectivity
164. 6:50module
165. 6:51and from static routes in the remote
166. 6:54access and vpn module
167. 6:57now this figure here shows a sample
168. 7:00enterprise network
169. 7:01with the distribution points throughout
170. 7:04okay so in this example some remote
171. 7:08sites
172. 7:09requires connectivity to the server form
173. 7:13okay therefore one way redistribution is
174. 7:17performed
175. 7:18to inject routes from these remote sites
176. 7:22into the campus core okay
177. 7:26now some remote sites requires
178. 7:28connectivity to the entire network
179. 7:31this is provided by a two-way
180. 7:33redistribution otherwise
181. 7:35static routes would have to be
182. 7:36configured on the campus score
183. 7:40now the building distribution layer
184. 7:42propagates only on the default
185. 7:44route down to the building access layer
186. 7:46whereas the building access layer
187. 7:48advertises
188. 7:49its own subnets to the building
189. 7:51distribution layer
190. 7:53the distribution might also be necessary
191. 7:56in the remote access and vpn and
192. 7:59internet connectivity modules
193. 8:02okay now for a remote access and vpn
194. 8:05module with static routing
195. 8:07static routes are injected into the
196. 8:09campus core
197. 8:11okay now in the opposite direction
198. 8:15default routes provide connectivity for
199. 8:17remote users
200. 8:20in internet connectivity module with
201. 8:23only one exit
202. 8:24point the exit point is the default
203. 8:27route
204. 8:28to traffic the standard internet and
205. 8:31is propagated through the core routing
206. 8:33protocol
207. 8:35now if multiple exit points towards the
208. 8:38multiple isps exist
209. 8:41bgp provides internet connectivity
210. 8:44and the distribution can be used
211. 8:51route filtering so as mentioned route
212. 8:54filtering might be required
213. 8:56when redistributing routes route
214. 8:58filtering prevents
215. 9:00the advertisement or acceptance of
216. 9:02certain routes
217. 9:03through the routing domain so filtering
218. 9:06can be configured as follows
219. 9:08so on a routing domain boundary where
220. 9:11distribution occurs
221. 9:13within the routing domain to isolate
222. 9:15some parts of the network from other
223. 9:17parts
224. 9:18to limit routing traffic from untrusted
225. 9:21external domains
226. 9:23filtering is used with router
227. 9:25distribution
228. 9:26primarily to prevent suboptimal
229. 9:29routing and routing loops that might
230. 9:32occur
231. 9:33when routes are redistributed at
232. 9:35multiple redistribution points
233. 9:38so route filtering is also used to
234. 9:40prevent
235. 9:41routes about certain networks such as
236. 9:44private ip address space from being sent
237. 9:49or to received from remote sites
238. 9:55okay so the next one would be route
239. 9:57summarization
240. 9:59so route summarization which is also
241. 10:01called route
242. 10:02aggregation or supernatant so
243. 10:05in route summarization a single summary
244. 10:08address in the routing table represents
245. 10:10a set of routes
246. 10:13summarization reduces the routing update
247. 10:15traffic
248. 10:16the number of routes in the routing
249. 10:18table and the overall router overhead
250. 10:22in the router receiving the routes now
251. 10:25what are the benefits
252. 10:26of route summarization so a large flat
253. 10:30network is not scalable because routing
254. 10:32traffic consumes considerable
255. 10:34network resources when a network change
256. 10:37occurs
257. 10:38it is propagated throughout the network
258. 10:40which requires
259. 10:41processing time for route
260. 10:43reconfiguration and bandwidth
261. 10:46to propagate routing objects
262. 10:52so for route summarization so for
263. 10:55example
264. 10:56from our given topology here okay
265. 11:00so you've got the san francisco campus
266. 11:02the denver region and the houston region
267. 11:05so route summarization should be on a
268. 11:09router
269. 11:09or the edge router connected to
270. 11:13other network in this case this
271. 11:16router here connects the denver region
272. 11:20to the san francisco campus so therefore
273. 11:22this is
274. 11:23where we perform route summarization
275. 11:26so instead of propagating all the
276. 11:28networks within the denver region
277. 11:31we just have to summarize this uh ip
278. 11:33addresses here and come up with a single
279. 11:37172.16.8.0 22
280. 11:39to represent the entire network on the
281. 11:42denver
282. 11:42region so you could do the same for
283. 11:45houston
284. 11:46going to san francisco
285. 11:52so recommended practice summarize at the
286. 11:55distribution
287. 11:56layer so it is recommended practice to
288. 11:59configure summarization in a large
289. 12:01network
290. 12:02from the distribution layer towards the
291. 12:04core as
292. 12:05illustrated on this figure okay
293. 12:09now the distribution layer should
294. 12:12summarize all networks
295. 12:14on all interfaces towards the campus
296. 12:16core
297. 12:17one connectivity and remote access
298. 12:19points should be summarized
299. 12:21towards the core also for example
300. 12:25remote subnets could be summarized into
301. 12:27major networks
302. 12:29and only those major networks would be
303. 12:31advertised to the core
304. 12:34now implementing summarization at the
305. 12:36distribution layer optimizes the
306. 12:38convergence process
307. 12:39for example if a link to an access layer
308. 12:43device
309. 12:44goes down return traffic to that device
310. 12:47is
311. 12:48dropped at the distribution layer until
312. 12:51the routing protocol converges
313. 12:53so summarizing also limits the number of
314. 12:57peers
315. 12:58that an eigrp router must query
316. 13:02or the number of lsas that an ospf
317. 13:05router must process
318. 13:06which also reduces the convergence time
319. 13:12recommended practice passive interfaces
320. 13:15for igp at access layer
321. 13:18so another recommended practice is to
322. 13:20limit unnecessary peering across the
323. 13:23access layer
324. 13:24now from the figure the distribution
325. 13:27multi
326. 13:27uh layer switches are directly connected
327. 13:30to each other
328. 13:31okay and are also interconnected with
329. 13:34three axis layer switches
330. 13:37each having four vlans for instance
331. 13:40now by default the distribution layer
332. 13:42devices send
333. 13:44routing updates and attempt to pair with
334. 13:47remote distribution layer devices across
335. 13:49the links
336. 13:50from the access switches on every vlan
337. 13:53now having the distribution switches
338. 13:56form neighbor relationship
339. 13:58over these 12 axis layer connections
340. 14:02provides no benefit and wastes of
341. 14:04resources
342. 14:06including cpu processing time and memory
343. 14:08so therefore
344. 14:10the interfaces on the distribution layer
345. 14:13devices towards the access layer
346. 14:16should be configured as passive
347. 14:18interface
348. 14:19under the routing protocol configuration
349. 14:22okay
350. 14:22this suppresses the advertisements of
351. 14:25routing updates
352. 14:26for that routing protocol on those
353. 14:28interfaces
354. 14:34now summarizing designing a routing
355. 14:36protocol deployment
356. 14:39large networks may implement multiple
357. 14:42protocols for different modules on the
358. 14:44cisco enterprise architecture
359. 14:47advanced routing features such as
360. 14:49distribution
361. 14:51filtering and summarization allows
362. 14:54multiple routing protocols to co-exist
363. 14:56and provide
364. 14:57greater scalability the distribution
365. 15:01between different routing protocols
366. 15:02passes routing knowledge from one
367. 15:05protocol to another
368. 15:07route filtering prevents advertisement
369. 15:10of certain routes through the routing
370. 15:13domain
371. 15:14route summarization and an ip hierarchy
372. 15:18reduce
373. 15:18routing traffic and unnecessary route
374. 15:21recomputation
375. 15:25so ip addressing and routing review so
376. 15:29that includes defining the ip addressing
377. 15:31requirements
378. 15:32develop hierarchical ip addressing plan
379. 15:36use private addresses inside or within
380. 15:39the organization
381. 15:41use public addresses facing the internet
382. 15:45use not or path for translation as
383. 15:48needed
384. 15:50develop a plan for deploying date gp and
385. 15:52dns
386. 15:54use eigrp or ospf based on
387. 15:57organizational requirements
388. 16:00and implement recommended practices
389. 16:03including redistribution
390. 16:05filtering and summarization
391. 16:09now let us summarize the module
392. 16:13we've talked about designing ip
393. 16:16addressing
394. 16:16and selecting routing protocols ip
395. 16:20address structure and ip
396. 16:22address types have a large impact
397. 16:26on the address plan for both ipv4 and
398. 16:28ipv6
399. 16:30eigrp and ospf are the recommended igps
400. 16:35for the enterprise advanced routing
401. 16:38features such as redistribution
402. 16:40filtering and summarization supports
403. 16:44scalability and multiple routing
404. 16:46protocols
405. 16:49that's the end of this video lecture see
406. 16:52you on the next video have a great day