Log shipping
takes time to back up the log, copy it to the standby server, and restore it.
??? Less potential data loss: Depending on the type and the configuration, replication
could theoretically prevent data loss altogether. Log shipping guarantees some
amount of loss of data. Of course, if the log backup includes no changes to the
data or structure of the database, nothing is lost.
??? Multiple, active standby servers: Log shipping provides for multiple standby
servers, but these servers can be used as read-only reporting databases at best.
With replication, the other standby servers are fully accessible.
It is this final point about accessible standby servers that begins my counterargument.
If you need to have fully accessible multiple servers (meaning available for insert,
update, and delete operations), then you have a distributed system. Data is being submitted
or altered from multiple points. So what exactly is the object of disaster recovery?
Each individual node, or the entire distributed system?
The primary purpose of replication is to supply pieces of information to or from a
centralized location.
Pages:
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286