This document compares all-flash storage solutions from NetApp and Pure Storage. It finds that NetApp has stronger performance capabilities, supported by its scale-out architecture and Data ONTAP innovations. However, Pure Storage leads in storage efficiency through its deduplication and compression techniques. While both vendors offer enterprise-grade reliability, NetApp provides more robust data management services and flexibility in scaling storage capacity and performance over time. The document recommends that NetApp focus on medium and large enterprises by emphasizing its strengths in scalability, performance, and data services management.
3. INFLUENTIAL TRENDS
Flash brings high performance to data services
Storage efficiency technologies
The newer application workloads are driven by mobile computing, social
media, big data/analytics and cloud computing
These associated higher workloads require storage performance (IOPS,
latency, throughput, consistency) that HDDs alone cannot cost effectively
meet
Data is expected to grow at 44% CAGR, so the ‘managing growth’ problem
will not let up soon. Flash is needed to maintain balance as infrastructure
density increases.
4. ECONOMICS
SSD Prices per GB have steadily
fallen since 2013, from about $2
to 69 cents while HDD’s remain
constant
FLASH AT SCALE
Flash at scale delivers significant secondary economic benefits
Fewer devices, reduced energy/floor space consumption, fewer
servers, lower s/w licensing costs
5. AFA MARKET – REVENUE & FORECAST
WorldwideAFA segment reached $1.3 billion in 2014
Market expected to grow 5 times in revenue by 2019
AFAs will dominate primary storage by 2018-2019
AFAVendors pursuing lower entry price points
Shorten sales cycles, newer customers
REVENUE ($M) SHARE (%)
EMC 112.3 22.6
Pure Storage 90.9 18.3
IBM 82.9 16.7
NetApp 45 9.1
SolidFire 35.6 7.2
Nimbus Data 34.3 6.9
Other 95.3 19.2
Total 496.3 100
6. FUTURE OUTLOOK
Mixed workload consolidation is the competitive
background for AFAs for 2015 and beyond
Independent software vendors (ISV) will develop
primary storage applications assuming all flash
configurations
VM-level storage management will become a way
of life by 2017
8. FOREWORD
The two products under comparison come from firms
whose competency and outlook in the AFA market is
different
Pure Storage’s competency lies in software (storage
efficiency) and reducing effective cost
NetApp’s competency is hardware, performance and
enterprise class data services
Ultimately, it is the application workload that governs
which solutions you should consider
9. TARGET MARKETS
The Pure Storage FA-405, FA-420 and FA-450 are
designed to address the small business and remote
offices, mid-market and high-end enterprise
markets respectively
Netapp All Flash FAS is targeted for enterprise
customers who want an all-flash system for
performance while maintaining their familiar
storage management and data protection features
10. PRODUCT ARCHITECTURE
NETAPP PURE STORAGE
Scale out approach - Is a cluster of servers
(nodes) with a single point of management
Scale up approach – One stand alone server
deployed. (Moore’s law allows for
replacement of the stand alone server in the
tech. refresh cycle. *How long now?)
COMMON GOALSOF EACH APPROACH
Scale performance, capacity or both non-disruptively and without performance
degradation.
Others include affordability, interoperability and operational simplicity
Regarded as the storage of the future
Mission critical workloads can run on 2 or
more servers to preserve availability
Suitable when a large number of smaller
nodes are needed, perhaps for a web server
farm or a server cluster where physically
redundant hosts are required
Suitable for major virtual server
consolidation
Works only if resources to a virtual machine
are carefully allocated
Potential for scale-up server failures and
work disruptions is real
*“I pick 2020 as the earliest date we could call [Moore's law] dead”, Bob Colwell (Intel)
11. PERFORMANCE
Performance is one of the most important metrics for
embracing All Flash Storage.
Impacts your business by increasing efficiency and
reducing server and licensing costs by 50%
Key factors are
IOPS (Higher the better)
Latency/Response time (Lower the better, <1ms)
Consistency
12. NETAPP WINS IN PERFORMANCE
NETAPP (STRENGTH) PURE STORAGE (WEAKNESS)
Scale out performance Use commodity SLC flash drives as
“NVRAM” to journal data, so maximum
theoretical write performance of only
150,000 4K IOPS*
685,281 IOPS, $2.68 IOPS and 1.23
millisecond response time at 100% load
(Variability in latency is low too)
The 512B granularity approach shall lead to
fragmentation over time.
And since in flash, sequential IO is faster
than random, when the system would have
to hop around (random read) to read in
512B blocks, performance dips
NetApp’s Data ONTAP FlashEssentials
innovations increase flash performance and
efficiency
500,000 overall IOPS sustained (@ 4KB)
200,000 write IOPS sustained (@ 4KB)*
Pure Storage’s ACTIVE/ACTIVE Controller Message (P.S. – Similar to
ACTIVE/PASSIVE)
Understand your requirements to determine if you need high performance or consistent
performance
13. STORAGE EFFICIENCY
Second most important metric enabling more data to be
stored in less space
Storage efficiency significantly reduces your cost of
ownership and effective cost
Areas of study include
Deduplication (Inline and cluster wide is desirable)
Compression (Inline is desirable)
In-line data reduction
Thin provisioning
14. PURE STORAGE WINS
Pure Storage is the leader in storage efficiency with it’s
differentiator being deduplication and compression
NETAPP’S WEAKNESS
The new AFF offers partial inline deduplication with
actual byte-to-byte comparison being post process
NetApp’s deduplication is volume based only
Data ONTAP lacks global deduplication across
LUNs which means that 250 LUNs can have 250
copies of same data adding up to a high cost flash
array
15. CAPACITY (NETAPP WINS)
NetApp’sAFF offers upto 384TB in raw capacity
and 1565.3 in effective capacity (Single HA Pair)
NAS Scale out (12 HA Pair) can offer 4.6 PB
SAN Scale out (4 HA Pair) can offer 1.5 PB
Pure Storage’s //m Series offers only upto 136TB in
raw capacity and 400TB in effective capacity but
with high storage efficiency
16. APPLICATION PERFORMANCE + DATA SERVICES
While Pure Storage supports enterprise applications
likeVDI, Oracle etc., superiorVDI application
performance is achieved on NetApp.
For databases however, Pure Storage wins
Data services like Replication, Encryption etc. are
refined at NetApp
17. APPLICATION PERFORMANCE + DATA SERVICES
NETAPP PURE STORAGE LEADS
Converged
Infrastructure
AllFlash FlexPod
available forVDI
environments
Hybrid Array FlexPod
available too
Flash Stack CI
announced Dec 2014
NETAPP
ORACLE & Microsoft
SQL
Supported Supported - It
requires 5-10x less
flash than hyper
converged approach
PURE STORAGE
VMWare (VDI) &Citrix Supported - Storage
costs as low as $55/
desktop with Horizon
View
Supported - Markets
only against disks
NETAPP
Openstack Deeply Integrated Recently added (2014) NETAPP
Data Services
Snapshots,
encryption,
replication,
Integration
Supported (MATURE) Supported and free in
Purity 4.0
(FlashProtect &
FlashRecover
(Launched 2014))
NETAPP
18. FLEXIBILITY AND SCALABILITY(NETAPP WINS)
How confident are you about your future data architectures?
Flexibility - NetApp’s FAS allows for a lot of flexibility by allowing
data to be moved from flash to disk to cloud.
NetApp offers both SAN & NAS architectures, while Pure Storage
offers SAN only
Scalability - A Pure system uses a whopping 384 GB of RAM to
deliver 120TB
Given that DRAM prices are dropping at a much slower pace than
flash, Pure is hitting the limits of commodity server design
19. QUALITY OF SERVICE (NETAPP WINS)
Built-in QoS lets you set performance limits on files,
volumes, LUNs, or an entire storage virtual machine
(SVM) to give priority to your most important
workloads.
RELIABILITY (STALEMATE)
Both companies boast >99.999% reliability across all systems deployed
Pure Storage - Proprietary RAID 3D is deployed
NetApp – RAID 6 or RAID 4 or combination of prior with RAID 1 is deployed
20. PURE STORAGE – STRENGTHS (INHERENT & TRANSIENT)
Cost efficient performance driven by deduplication and compression driving down
effective cost
Creative pricing and guarantee programs around controller upgrades, SSD
warranties and maintenance pricing
Customer trust by demonstrating it is reliable, stable and easy to work with
Delivers capable data management services and proven data reduction
implementation while based on consumer-grade SSD technology
Accelerate the I/O performance for relational databases, virtual desktop
infrastructure (VDI), and virtual server infrastructure and clouds.
Others (TRANSIENT) – Gartner Magic Quadrant Leadership status
21. PURE STORAGE - WEAKNESSES
Performance
Lack of enterprise class data management and
application features, NetApp standard
Pure Storage’s FA-400 series product has limited
scalability (read high RAM/Storage ratio) when
compared to competitive products.
22. NETAPP - STRENGTHS
Robust performance
Leading data management services and ecosystem
Scalable – All flash systems will run out of capacity
before performance levels decrease
All-Flash FAS is available as part of the FlexPod
reference architecture
Alliances (Cisco,VMWare etc.)
23. NETAPP - WEAKNESSES
Lack of data reduction capabilities limits the appeal
of FAS Series in server virtualization, virtual desktop
infrastructure (VDI) and online transaction
processing (OLTP) consolidation use cases.
Creative Marketing capabilities vis-à-vis Pure
Storage
24. PROMOTION (TRANSIENT – Info only)
Pure Storage introduced the ForeverFlash initiative,
which combines the ability to non-disruptively scale
performance or capacity with a maintenance program
that promises no hikes in annual fees.
When users extend support for another period, Pure
provides new controllers at no cost.
Caveat
Pure must continue to excel at R&D and deliver upgrades
from engineering
With majority of sales from channel, how would the finances
work out?
25. PRICING & STANDARD WARRANTY – (STALEMATE -
PRICE/SERVICE WARS)
FACT - Flash performance is so good that customers want to stay on one platform for as long as possible
PURE STORAGE
FlashArray 405 & 450
List prices range from sub-$100K to multi-$100K.
Pure's raw flash is the most expensive at $12.03/GB
Standard warranty is 3 years from initial purchase.
Maintenance and support are offered in 1, 3 and 5 year plans, each with 4-hour or next business day (NBD)
options (both options are 24x7).
Pricing for support is based on percentage of list price:
1 year 4-hour is 8%, NBD is 6%
3 year 4-hour is 20%, NBD is 15%
5 year 4-hour is 33%, NBD is 26%
NETAPP
Starting street price is $25,000
AFF8000 is now priced at the same $5 per raw GB
Support flat maintenance for up to seven years versus the typical three to five years, so there's no penalty for
staying on the platform for longer times
26. GAPS IN NETAPP’S OFFERING
Storage efficiency is an industry requirement and
NetApp is lagging.
NetApp lacks ease of use for the SME target
market. Pure Storage achieves this well.
Training and professional services are required for
most of NetApp’s deployments
27. RECOMMENDATIONS
LONGTERM
Product Enhancements
SHORTTERM – SAY “Capacity is free, performance costs”
Focus on the medium & high-end enterprises targeted by Pure
Storage’s FA-450 only.They are more inclined to need performance,
data services, application integration etc.
Leverage and build a compelling case for scalability, performance
and data management services (NetApp’s strengths) during the
pitch
Storage is a long term investment, caution the customer to not rush
into the decision based on cost (Pure Storage’s strength)
“Moore’s law will be dead in a decade”, Bob Colwell (Intel), marking
an end to scale up architecture (Pure Storage)
Discuss NetApp’s whole range of offerings
28. RECOMMENDATIONS
Partner with ISW’s/Alliances to offer discounts on licenses
Attack Pure Storage’s hardware/weakness & marketing (Consumer
grade SSDs, Performance, 50% CPU utilization etc.)
Tackle the prospect’s H/W and S/W needs separately (Review and
refine product offerings, Unbundle premium offerings)
Focus on customer priority - $/GB or $/IOPS ?