2. ABSTRACT
• “The zettabyte file system” a file system that provides strong
integrity guarantees , simple administration and immense capacity.
• This architecture includes pooled storage ,a transactional copy on
write model .
• It reduces implementation complexity allows new performance
improvements and provides several useful new features.
3. INTRODUCTION
• Upon hearing about on ZFS some appear to be genuinely surprised and ask
“aren’t local file systems a solved problem?”.
• We can deduce that the speaker has probably never lost important file and run
out of space on partion , struggled with a volume manager.
• so we begin to design of ZFS with the goals of strong integrity, simple
administration and immense capacity.
• ZFS turned out to be a simpler to implement than many recent file systems ,
which are hopeful to sign for the long term viability of our design.
4. DESIGN PRINCIPLES
• Some of design principles are:
simple administration
dynamic file system size
immense capacity
error detection and correction
5. SIMPLE ADMINISTRATION
• On the most systems ,creating a logical device , and creating a new fille system is
painful and slow operations
• Their isn’t much pressure to simplify speed up these kind of administrative tasks
cause they are relatively uncommon and only performed by the administrator.
• The administrator should be able to add more storage space to an existing file
without unmounting ,locking , or otherwise interrupting service on that file
system.
• Administrating’s storage should be fast, simple and difficult to screw up
6. DYNAMIC FILE SYSTEM SIZE
• If a file system can only use space from its partition, the system should predict
the max future size of each file system at the time of creation
• Most of the disk space may be unused , but if no room is left for a new partition ,
no new file system is created.
• Once the storage used by the a file system can grow and shrink dynamically as
users add and remove data .
• Instead , structures such as inodes square should be allocated and freed
dynamically as files are created and deleted.
7. IMMENSE CAPACITY
• File systems want to handle 16EB of data need more than bigger block address.
• They also need scalable algorithms for directory lookup , meta data allocation
and routine operations.
• The on-disk format deserves special attention to make sure it wont preclude
scaling in some fundamental way.
• To maintain disk consistency it shouldn’t depend upon the fsck.
8. ERROR DETECTION AND CORRECTION
• Bugs in disk controller firmware can result in variety of errors,
• Including misdirected write and mis directed reads and phantom writes.
• The file system should validate data read in from disk in order to detect many
kinds of filesystem corruption.
• The file system should automatically correct corruption ,if possible ,by writing the
the correct block back to the disk.
11. CONCLUSION
• Current file system still suffer from a variety of
problems:
• complicated administration
• poor data integration
• limited capacity
• Our implementation is relatively simple , especially
considering that we incoperate all the functionality of a
volume manager
• ZFS sets an entirely new standard for data integrity
,capacity ,and ease of administration in the file systems.
