For Project Enquiry +91 9962 067 067

REDUCING FRAGMENTATION FOR IN-LINE DEDUPLICATION BACKUP STORAGE VIA EXPLOITING BACKUP HISTORY AND CACHE KNOWLEDGE

ABSTRACT

In this paper, backup systems, the chunks of each backup are physically scattered after de-duplication, which causes a challenging fragmentation problem. Observe that the fragmentation comes into sparse and out-of-order containers. The sparse container decreases restore performance and garbage collection efficiency, while the out-of-order container decreases restore performance if the restore cache is small. In order to reduce the fragmentation, we propose History-Aware Rewriting algorithm and Cache-Aware Filter. History-Aware Rewriting exploits historical information in backup systems to accurately identify and reduce sparse containers, and Cache-Aware Filter exploits restore cache knowledge to identify the out-of-order containers that hurt restore performance. Cache-Aware Filter efficiently complements History-Aware Rewriting in datasets where out-of-order containers are dominant. To reduce the metadata overhead of the garbage collection, we further propose a Container-Marker Algorithm to identify valid containers instead of valid chunks. Our extensive experimental results from real-world datasets show History-Aware Rewriting significantly improves the restore performance by 2.84-175.36x at a cost of only rewriting 0.5-2.03% data.

EXISTING SYSTEM	PROPOSED SYSTEM
EXISTING CONCEPT:- In the existing system, Existing garbage collection solutions first identify valid chunks and the containers holding only a few valid chunks (i.e., reference management). Then, a merging operation is required to copy the valid chunks in the identified containers to new containers. Finally, the identified containers are reclaimed. Unfortunately, the metadata space overhead of reference management is proportional to the number of chunks, and the merging operation is the most time consuming phase in garbage collection.	PROPOSED CONCEPT:- In the proposed system fragmentation problem in de-duplication systems has received many attentions. IDedup eliminates sequential and duplicate chunks in the context of primary storage systems. Nam et al. propose a quantitative metric to measure the fragmentation level of de-duplication systems, and a selective de-duplication scheme for backup workloads. SAR stores hot chunks in SSD to accelerate reads. RevDedup employs a hybrid inline and out-of-line de-duplication scheme to improve restore performance of latest backups.
EXISTING ALGORITHM:- Rewriting algorithm.	PROPOSED ALGORITHM:- History-Aware Rewriting Algorithm
ALGORITHM DEFNITION:- Rewriting algorithm, which is a tradeoff between de-duplication ratio (the size of the non-de-duplicated data divided by that of the de-duplicated data) and restore performance. These approaches buffer a small part of the on-going backup stream, and identify the fragmented chunks within the buffer.	ALGORITHM DEFNITION:- HAR rewrites all duplicate chunks whose container IDs exist in Sinherited. Additionally, HAR maintains an in-memory structure, Semerging (included in collected info in Figure 3), to monitor the utilizations of all the containers referenced by the backup.
DRAWBACKS:- Rewriting algorithms not able to achieve best restore performance with a larger restore cache.	ADVANTAGES:- HAR is able to achieve best restore performance with a larger restore cache. The hybrid scheme is significantly better than other rewriting algorithms