mạng lưới (WAN), khả năng trì hoãn khi truyền tải những thông điệp cam kết hơn nữa có thể tăng thời gian đáp ứng I / O và ảnh hưởng đến hiệu suất của phần mềm dựa trên phản ứng I / O nhanh chóng. Như vậy, mạng chất lượng cao với băng thông cao, tổn thất thấp, và sự chậm trễ chút là bắt buộc. | Replication Strategies 241 network WAN the possibility of delay when transmitting the commit messages can further increase I O response times and affect the performance of software that relies on prompt I O response. Thus high-quality networking with high bandwidth low loss and little delay is required. Some synchronous solutions can throttle the rate of updates according to network load. Furthermore the frequency and size of data updates sent over the WAN can consume network bandwidth 11 . Asynchronous Data Transfer Asynchronous replication involves performing updates to data at a primary storage site and then sending and applying the same updates at a later time to other mirrored sites. The updates are sent and applied to the mirrored sites at various time intervals. This is why asynchronous data transfer is sometimes referred to as shadowing. With asynchronous data transfer updates committed at the primary site since the last PIT of issue to the mirrored site could be lost if there is a failure at the primary site. Thus the mirrored site will not see these transactions unless they are somehow recovered at the primary site. Thus some consideration is required in specifying the right interval to meet the RPO. The smaller the RPO the more frequent are the update intervals to the mirrored database. Asynchronous replication is illustrated in Figure . A read write request is immediately committed at the primary disk system. A copy of that same request is sent at a later time to the mirrored system. The primary system does not have to wait for the mirrored system to commit the update thereby improving I O response times at the primary system. The local disk subsystem at the primary site immediately informs the application that the update has been made usually within a few milliseconds. As the remote subsystem is updated at a later time the application performance at the primary site is unaffected. As stated earlier a caveat with asynchronous .
