Cohesity SnapTree Solution Brief

Download as pdf or txt
Download as pdf or txt
You are on page 1of 2

Cohesity SnapTree TM

Executive Summary:
The 24/7 nature of today’s enterprise requires that Recovery Time Objectives
(RTO) and Recovery Point Objectives (RPO) are reduced to a minimum. Cohesity
Data Protect is built on patented Cohesity SnapTree technology to deliver a
robust data protection strategy that delivers (i) Protection from accidentally
deleted files, application crashes, data corruption (ii) Data retention for long
Cohesity SnapTree
periods of time to satisfy compliance and regulatory requirements (iii) Low RPO Benefits
& RTO with a distributed web scale platform for secondary storage infrastructure.
•A
 chieve faster backup and restore
Challenges with current Snapshot technologies: Data protection in legacy with reduction in RTO and RPO
storage solutions leverage copy-on-write/redirect-on-write snapshot technology windows by over 90%
to create copies of data. These snapshots of a file system at a particular given
point in time form a chain, tracking the changes made to a set of data and
•O
 btain pay as you grow scaling
form the basis for organizing and storing copies of data. Every time a change
that ensures maximum data
is captured, a new link is added to the chain. As these chains grow with each
reduction while backups are
and every snapshot, the time it takes to retrieve data on a given request grows
readily available for instant
because the system must re-link the chain to access that data. This process is
recovery and clone capability
highly inefficient and is unable to cope with the growing demands of achieving
shorter RPO and RTO windows.
•L
 ower Total cost of ownership
Cohesity SnapTree™ Technology (TCO) by limiting storage growth
from multiple data copies
Cohesity has brought innovation to data protection through the SnapTree
technology. SnapTree is a ‘Distributed-Redirect-on-Write’ (DROW) snapshot
mechanism that provides speed and scalability in addition to the inherent
benefits of RoW snapshot. The design is optimized for write performance so
any changes are redirected to new blocks. Additionally, all nodes participate in
this process thereby leveraging the scalability elements of the Cohesity cluster.
SnapTree is available as part of Cohesity DataPlatform.

Comparison between Traditional snapshots and Cohesity SnapTree

Traditional snapshots Cohesity SnapTree

Read/write Degrades as the number of No change as the number of


performance traverses to reconstruct data traverses to reconstruct data from
increases linearly with time. any snapshot is fixed.

Number of Limited by performance impact Unlimited because data change


snapshots of frequent snapshots particularly does not result in accesses needed
with write intensive workloads. to retrieve the data. Moreover, the
High data change rate will mean snapshots are distributed across all
high IO (reads/writes) overhead nodes in a cluster, hence the term
on any changes/updates after DROW snapshots.
snapshots are taken.

RPO Order of hours to days. Order of minutes.

RTO Since snapshots are not fully SnapTree provides fully hydrated
hydrated, RTO times are high with snapshots to recover quickly to a
traditional snapshots. certain point in time.

Instantaneous with SnapTree.


Cloning for test/
One Platform.
Several vendors consolidate
dev purpose changes to build out periodic full
copies. This can take significant
time due to traversing the entire Infinite Possibilities.
snapshot chain.

Storage Inefficient due to periodic fulls Efficient by eliminating the need to


consumption required to consolidate data do periodic fulls.
changes over time.
Cohesity’s patented SnapTree technology allows Datafile reconstruction using Conventional Snapshot images

businesses to take a large number of snapshots at


any time interval with uncapped retention policies, S0 S1 S2 S3 Sn Unbounded
accumulated
without ever affecting performance or consuming traverses (n) to
reconstruct
additional space. Cohesity’s DROW implementation A B C0 C1 C2 C3 Cn A, B, Cn
keeps track of changes by writing the changed data
to new blocks.
Datafile reconstruction using Cohesity SnapTree images

SnapTree creates a tree of pointers that limits the


S0 S1 S2 S3 Sn
number of traverses it takes to retrieve blocks of Always fixed traverses
(2 in this example) to
data, regardless of the number of snapshots that reconstruct
A, B, Cn
have been taken. SnapTree uses a B+ tree data
structure (Fig 1) such that access to any point in
the tree takes a fixed number no matter how many A B C0 C1 C2 C3 Cn
snapshots there are, without having to rebuild any
chain linkage. Time

Fig 1. Cohesity SnapTree technology helps create snapshots without


incurring the recovery penalty of traversing the entire snapshot chain
seen in traditional snapshot architecture.

Because SnapTree is implemented on a distributed Cluster


file system (Fig 2), every node sees the same
Node 1 ... Node 2 ................ Node n
nested structure of the chain with a fixed depth
independent of where the actual data is stored in
the cluster. Keeping the snapshots fully hydrated
improves the recovery times of any snapshot from
t0 to tn because it does not incur the time penalty of
S0 S1 S2 S3 Sn
traversing the entire chain of changes. Each of these Every node in the
cluster sees fixed
snapshot clones is fully hydrated so that businesses traverses (2 in this example)
to reconstruct A, B, Cn
can achieve fast RTO and near-continuous RPO
objectives. SnapTree is available as part of Cohesity
DataPlatform. A B C0 C1 C2 C3 Cn

Time

Fig 2. Distributed-Redirect-on-Write’ (DROW) snapshot ensures that


every node sees the same nested structure of the SnapTree

Address 451 El Camino Real, Santa Clara, CA 95050 @cohesity


Email [email protected] www.cohesity.com ©2016 Cohesity. All Rights Reserved.

You might also like