业务永续-数据中心的业务连续性规划与设计l.pptVIP

* * * * * * Now the discussion switches to resilency systems availability. We lead with this great report from Yankee Group that shows that AIX systems have lless than half the down time of HP-UX and Solaris systems and much less than 10% of the downtime of Windows systems. No wonder users select UNIX for their critical workloads * * * PowerHA/XD – Disaster Recovery GLVM asynchronous mode* Native AIX IP based replication Async mode for geographic dispersion PowerHA/XD – Disaster Recovery Support for Global Mirror for the SVC Storage Volume Controller (SVC) Async replication for geographic dispersion WebSMIT gateway – Cluster Management Manage multiple clusters from a single GUI interface No footprint or overhead on managed clusters Deploy from a server outside of the cluster domain IPV6 for PowerHA cluster client - Security IPV6 is the follow on technology to IPV4 Client access to PowerHA cluster enabled for IPV6 option * Locating the datacenters further apart - say for example on different parts of the power grid or different parts of a fault line - requires networking which can span greater distances. GLVM can mirror data across any TCP/IP capable network connection. In this setup, a standard LVM volume group is configured with mirroring, but some of the mirrors are pseudo devices which copy the data across the network to the backup site. This operation is transparent to the application, though there will be some performance impact as the distance between sites increases For clients with multiple datacenters in the same metropolitan area, hardware based replication like Metro-Mirror (formerly PPRC) can be used to keep the data in sync. Metro-Mirror requires dedicated networking between the sites which limits the range to about 100km without extenders. There are 2 levels at which Metro-Mirror can be implemented SAN Volume Controller or SVC mirrors data through the SVC controller at each site. ESS – or Shark subsystem mirrors data at the disk level. Bot