svc基于异构环境下的异构存储解决方案好文档.pptVIP

  • 0
  • 0
  • 约2.53万字
  • 约 52页
  • 2018-06-23 发布于湖北
  • 举报

svc基于异构环境下的异构存储解决方案好文档.ppt

svc基于异构环境下的异构存储解决方案好文档

* * * IBM offers a comprehensive line of tape storage options for System z9 and zSeries. Together, the tape drives and libraries are examples of the IBM focus on providing value to our z9 and open systems customers. self explanatory * * * * * The major difference between ProtecTIER and other standard VTLs is ProtecTIER’s unique patent pending factoring algorithm called HyperFactor. HyperFactor is the data de-duplication technology developed by Diligent Technologies that allows you to store a lot more backup data onto a smaller amount of disk. Here’s how HyperFactor works. As we build the slide, you’ll see that there are unique blocks of data in the repository. HyperFactor keeps track of this data with a memory resident index, like a table of contents, that can map the contents of a 1PB repository in only 4GB of memory. That 250,000:1 ratio between the repository and the index is a significant differentiator and has orders of magnitude greater granularity than anything in the market place. And as we build the slide, you’ll see a stream of data coming from one of the backup servers. This stream contains some data that already exists (as represented by the multi colored icons) and some data that’s new (as represented by the tan icons). HyperFactor is an inline data de-duplicator. As the data stream passes through the HyperFactor data de-duplication engine, HyperFactor will be looking for data that is “similar” to data it has stored before. PAUSE I use the word “similar” because it’s not identical, and that’s because part of the algorithm’s power is that it uses similarities instead of identicals to achieve unmatched performance. The most similar pattern in the repository is found with no I/O operations and then that data into brought to the server to do a computational compare and then store the delta. This is performed without impacting the search time, regardless of the repository size. PAUSE Because there is no I/O required, we’re actually performing a memor

文档评论(0)

1亿VIP精品文档

相关文档