06-09-2010, 07:14 PM
04-05-2011, 12:20 AM
Critical Review
This paper introduce the zfs system, a 128-bit file system. In the paper, they mainly introduce many features and design of the system. Such as simple administraion, pooled storage, etc. They also show there storage model and the intering storage pool alocator.At last they show the demo zfs in practise, and show there future work.
In my opinion, there are two intering points of the paper. One is the 128-bit file system, it seems that 128-bit store much more and more large of data than 64-bit. Just us we use IPV6 instead of IPV4.However, I still can not conclude that it will be the final file system. The development is always beyond our mind. And , it will be our system in the end, though we don't use it now.The other thing is pooled storage and the storage pool alocator. It is quite familar with the pool we used in database. By using this , we can put all the device in the storage pool, and be managed by storage pool alocator. If the storage is increased, the system can use it immediately.However, in the meantime, the zfs is not satisfied the goal in the paper. And the design is for improving the write speed. It may have a poor performance in random read. In conclusion , the zfs system is still a predictable system which will give some ideas of future 128-bit file system.
This paper introduce the zfs system, a 128-bit file system. In the paper, they mainly introduce many features and design of the system. Such as simple administraion, pooled storage, etc. They also show there storage model and the intering storage pool alocator.At last they show the demo zfs in practise, and show there future work.
In my opinion, there are two intering points of the paper. One is the 128-bit file system, it seems that 128-bit store much more and more large of data than 64-bit. Just us we use IPV6 instead of IPV4.However, I still can not conclude that it will be the final file system. The development is always beyond our mind. And , it will be our system in the end, though we don't use it now.The other thing is pooled storage and the storage pool alocator. It is quite familar with the pool we used in database. By using this , we can put all the device in the storage pool, and be managed by storage pool alocator. If the storage is increased, the system can use it immediately.However, in the meantime, the zfs is not satisfied the goal in the paper. And the design is for improving the write speed. It may have a poor performance in random read. In conclusion , the zfs system is still a predictable system which will give some ideas of future 128-bit file system.