Novel Approach: Clustered Cache Parallel NFS and
Implementing Across Wide Area Network: Disaster
Recovery Centers
Latesh Kumar K J
1
and R Lawrance
2
1
Department of CSE, Siddaganga Institute of Technology, Tumkur, Karnataka
2
Department of MCA, Ayya Nadar Janaki Ammal College, Sivakasi, Tamilnadu
Email: latesh.kj@hotmail.com, lawrancer@yahoo.com
Abstract— Bulk Data transfer is not a primary concern, network latency still impacts
performance over wide area network using the traditional data sharing protocols. Many
communications protocols, including common file sharing protocols such as network file
system (NFS) and common Internet file system (CIFS), were designed to work in a local
area network (LAN) environment where latency is extremely low. Even simple actions, such
as retrieving the attributes of a file, can require numerous round trips across the network.
In a WAN environment, each of those round-trips carries a latency penalty, dramatically
slowing operations; here is a new approach of using the CpNFS extension to the popular
NFS standard, to overcome the latency and bulk carriage across data center facility for
disaster management computing system and between primary and backup disaster recovery
centers.
Index Terms— CpNFS, NFS, CIFS, RAID, IETF
I. INTRODUCTION
Disaster center computing can have very different requirements than the more common and computationally
less demanding grid [9] computing. Although the size of the computational resources requested by a disaster
centre present its own difficulties, an onerous difficulty is satisfying the associated data requests to the
various clients. In addition to backing up very large datasets, sometimes the datasets are also be mirrored to
redirected disaster centre’s based on business requirement, which has unpredictable characteristics, must be
handled successfully or the whole process may be for naught. Conventionally, the dataset moves to the
premeditated disaster recovery resource base which redirect to another backup disaster recovery site
available, which also requires a move to some permanent home. An error at any point can lead to job
termination, with possible loss of huge data. These concerns have lead to great interest by both users and data
center management in NFS file systems because of the latency. A cluster pNFS [11] file system greatly
simplifies the process by allowing all I/O operations to occur as if to a local resource with new technology in
place by balancing the failover situations. Data centre computing is particularly amenable to this approach,
which generally transfers sufficiently large datasets such that latency concerns are not paramount. This new
approach is mainly intending to remove traditional performance bottlenecks, reduces the total cost of
© Elsevier, 2014
Proc. of Int. Conf. on Advances in Communication, Network, and Computing, CNC