Lab Validation Report

NetApp FAS8000 Series

Unified, Enterprise, Scale-out Storage

By Tony Palmer, Senior Lab Analyst r

August 2014

Introduction ...... 3 Background ...... 3 NetApp FAS8000 ...... 4 ESG Lab Validation ...... 6 Seamlessly Scalable Capacity and Performance ...... 6 Non-disruptive Availability and Maintenance ...... 12 Validating Cost-effective, Proven Storage Flexibility ...... 16 ESG Lab Validation Highlights ...... 21 Issues to Consider ...... 21 The Bigger Truth ...... 22 Appendix ...... 23

ESG Lab Reports The goal of ESG Lab reports is to educate IT professionals about data center technology products for companies of all types and sizes. ESG Lab reports are not meant to replace the evaluation process that should be conducted before making purchasing decisions, but rather to provide insight into these emerging technologies. Our objective is to go over some of the more valuable feature/functions of products, show how they can be used to solve real customer problems and identify any areas needing improvement. ESG Lab's expert third-party perspective is based on our own hands-on testing as well as on interviews with customers who use these products in production environments. This ESG Lab report was sponsored by NetApp.

All trademark names are property of their respective companies. Information contained in this publication has been obtained by sources The Enterprise Strategy Group (ESG) considers to be reliable but is not warranted by ESG. This publication may contain opinions of ESG, which are subject to change from time to time. This publication is copyrighted by The Enterprise Strategy Group, Inc. Any reproduction or redistribution of this publication, in whole or in part, whether in hard-copy format, electronically, or otherwise to persons not authorized to receive it, without the express consent of The Enterprise Strategy Group, Inc., is in violation of U.S. copyright law and will be subject to an action for civil damages and, if applicable, criminal prosecution. Should you have any questions, please contact ESG Client Relations at 508.482.0188.

© 2014 by The Enterprise Strategy Group, Inc. All Rights Reserved. Lab Validation: NetApp FAS8000 Series 3 Introduction This ESG Lab Validation Report presents the hands-on evaluation and testing results of the newly released NetApp FAS8000 with Data ONTAP 8.2. ESG Lab focused on key areas that make the FAS8000 an attractive offering for the midmarket and enterprise: enhanced capacity and performance, seamless scalability, non-disruptive availability, and cost-effective storage efficiency.

Background The demands on and for storage are increasing rapidly. To address data growth without interrupting business operations, rapid deployment of storage and IT resources becomes a function of scalability. ESG’s 2014 IT Spending Intentions Survey revealed that, as they have been since 2010, increased use of server virtualization, improved backup and recovery, and data growth management are all in the top ten most frequently cited 2014 IT priorities. The adoption of server virtualization is nearly ubiquitous among enterprise and midmarket organizations, and increased usage of the technology was identified by 32% of respondents as one of their most important IT priorities for 2014.1 Figure 1. Top-ten Most Important IT Priorities for 2014 Which of the following would you consider to be your organization’s most important IT priorities over the next 12 months? (Percent of respondents, N=562, ten responses accepted) Increased use of server virtualization 32% Information security initiatives 32% Improve data backup and recovery 29% Manage data growth 25% Desktop virtualization 24% Use cloud infrastructure services 23% Regulatory compliance initiatives 23% Major application deployments or upgrades 23% Business intelligence/data analytics initiatives 23% Data center consolidation 22%

0% 5% 10% 15% 20% 25% 30% 35% Source: Enterprise Strategy Group, 2014. ESG research also shows that midmarket organizations—those with less than 1,000 employees—share many of the same IT priorities as enterprise organizations. For the midmarket, data protection—in the form of improved data backup and recovery processes and/or business continuity/disaster recovery programs—accounted for the top-two responses. Also among the top-ten responses are increased use of server virtualization, managing data growth, and desktop virtualization. Traditional storage area networks (SANs) use a scale-up architecture, expanding capacity by adding disks to a monolithic controller architecture to scale capacity, and adding additional independent controllers to scale I/O. With more focus being placed on advanced capabilities such as data analytics, virtualization, and the transition to the cloud model, senior executives and line-of-business stakeholders are growing dependent on IT infrastructure. These key stakeholders are demanding that IT duplicate public cloud reliability and uptime, putting pressure on IT to develop non-disruptive operations and service level objectives.

1 Source: ESG Research Report, IT Spending Intentions Survey, February 2014.

© 2014 by The Enterprise Strategy Group, Inc. All Rights Reserved. Lab Validation: NetApp FAS8000 Series 4 NetApp FAS8000 The newly released NetApp FAS8000 series storage systems are designed to consolidate the FAS/V3200 and FAS/V6200 product lines into a single, enterprise scale-out storage line. The intent is to minimize overlap and simplify buying decisions for NetApp customers. The FAS8000 series is complemented by the newly released FAS2500 series storage systems which replace the FAS2200 series and provide a single scale-out storage solution for SMB and midmarket customers. Figure 2. The Comprehensive NetApp FAS Portfolio

NetApp FAS8000 enterprise storage systems are engineered specifically to address the demands of data-driven businesses using an integrated combination of high-performance hardware and scalable storage software. Leveraging the latest multi-core, multi-processor chipsets from Intel, the FAS8000 increases and enhances NVRAM to accelerate and optimize writes along with PCIe gen3 I/O architecture to maximize data throughput. Flexibility and expandability are provided by a dense modular package integrating the unified target adapter (UTA2) for 16G FC or 10Gb Ethernet communication capabilities on the same ASIC. The FAS8000 series supports active-active dual-controller high-availability in four configurations. A FAS 8020 HA pair scales to 2.88PB, while a FAS8040 HA pair scales to 4.32PB, a FAS8060 HA pair scales to 7.2PB and a FAS8080 EX HA pair scales up to 8.64PB and a maximum of 1,440 disk drives per controller pair. With up to 24 PCIe expansion slots, administrators can install UTA2, 1Gb and 10Gb Ethernet adapters, and 6Gb SAS adapters for disk expansion. FAS8000 can accelerate storage with up to 36TB of hybrid flash storage, improving throughput and lowering latency for predictable high performance. All flash FAS8000 configurations with up to 384TB of flash per HA pair provide the lowest latencies for maximum performance. NetApp FAS8000 leverages the latest generation of the Data ONTAP software in single controller, active-active controller, or clustered for high-availability configurations. Data ONTAP can cluster up to 24 nodes, leveraging the

© 2014 by The Enterprise Strategy Group, Inc. All Rights Reserved. Lab Validation: NetApp FAS8000 Series 5 combined I/O capability of all nodes for throughput and redundancy, ensuring enterprise accessibility to business- critical data. A single cluster can non-disruptively scale out to 432TB of hybrid flash and 103.7PB of capacity. A cluster of all-flash FAS8000s can scale out to 4.6PB of flash. New for the FAS8000 series is FlexArray storage virtualization. This feature enables the FAS8000 series controllers to pool storage provided by NetApp E-Series, EMC, HP, and Hitachi storage systems. FlexArray is enabled by license key on the FAS8000 series, eliminating the need for a separate hardware option. Using FlexArray, administrators can pool their heterogeneous storage systems, presenting a unified storage environment and enabling Data ONTAP storage functionality on these disparate storage systems. In addition, the heterogeneous storage environment can be managed by OnCommand, NetApp’s storage management console. The FAS8000 integrates the latest generation of the NetApp unified target adapter, known as UTA2. This advanced adapter implements two SFP+ ports. Each ASIC has two ports and these port pairs are field-selectable for either 16GFC or 10GbE. Customers can change the personality of UTA2 ports in software without having to swap out cards. Using the NetApp UTA2, administrators can support all block and file protocols using onboard ports in the FAS8000, configured exactly as needed. This simplifies planning and management because each FAS system can be precisely configured to match each organization’s physical plant. In addition to supporting multi-protocol storage (FC, FCoE, iSCSI, NFS, pNFS, CIFS/SMB, HTTP, and FTP), Data ONTAP, combined with the FAS8000, provides many advanced storage and storage management features. Data protection is provided by a suite of packages including SnapVault disk-to-disk backup, SnapManager application- and virtual machine-aware backup and cloning, FlexClone instant virtual copies of databases and virtual machines, and SnapMirror site-to-site mirroring for disaster recovery. OnCommand Balance for NetApp provides advanced analytics, enabling administrators to understand and tune the behavior of the storage architecture.

© 2014 by The Enterprise Strategy Group, Inc. All Rights Reserved. Lab Validation: NetApp FAS8000 Series 6 ESG Lab Validation ESG Lab performed hands-on evaluation and testing of the FAS8000 series at NetApp headquarters, in Sunnyvale, California. Testing was designed to examine the enhanced performance and seamless scalability of the FAS8000 platform, as well as the benefits of Clustered Data ONTAP, including non-disruptive operations and capacity efficiency. Results were analyzed and performance metrics were compared to the previous generation FAS platform, tested by ESG Lab in 2013.

Seamlessly Scalable Capacity and Performance The configuration used in ESG Lab testing is shown in Figure 3. Two physical servers were configured with VMware vSphere 5.1, with a total of 8 virtual machines. Each Physical Server was connected via dual UTA2 ports set to 10GbE personality to a NetApp clustered Data ONTAP 8.2.1 system comprised of two FAS8020 nodes, each with 12TB of capacity available for allocation. Microsoft Server 2008 R2 was installed as the Guest OS on each VM, and the Iometer workload generation utility was installed on each VM. The NetApp cluster was configured with two aggregates, one on each node. Each VM was allocated a dedicated LUN from within a volume on one of the two aggregates. These LUNs were used as the data drive for each VM and the target for each VM’s workload. Testing was conducted using the industry-standard Iometer test tool to both simulate workloads and create datasets. The applications simulated included Microsoft Exchange 2013, Microsoft SQL Server 2012, File Services, and disk-to-disk backup. Figure 3. The ESG Lab Test Bed

ESG Lab Testing In a way, storage system benchmark testing is like an analysis of the performance of a car. Specifications including horsepower and acceleration from zero to sixty are a good first pass indicator of a car’s performance. While specifications provide a good starting point, there are a variety of other factors that should be taken into consideration including the condition of the road, the skill of the driver and gas mileage ratings. Much like buying a car, a test drive with real-world application traffic is the best way to determine how a storage system will perform under real-world conditions. ESG Lab performance testing began with evaluation of low level IOPS and throughout characterization tests using the open source Iometer workload generation utility. I/Os per second, or IOPS, is a measure of the number of operations that a storage system can perform. When a system is able to move a lot of IOPS, it will tend to be able to service more applications and users in parallel. Throughput is an indicator of the available bandwidth and data-

© 2014 by The Enterprise Strategy Group, Inc. All Rights Reserved. Lab Validation: NetApp FAS8000 Series 7 moving capabilities of a storage system. Iometer profiles of 512KB sequential reads and 4KB random reads were used for this first pass analysis of the raw aggregate IOPS and throughput capabilities of the NetApp FAS8000 platform. As Figure 4 shows, the two-node FAS8020 cluster was able to drive 150,889 IOPS, more than 3x the number achieved by the two-node FAS3220 cluster in 2013. The FAS8020 cluster was able to support 1,839 MB/sec, nearly 2X the 974MB/sec throughput of the FAS3220 cluster, using the same clients and workload generation tool, connecting through the same number of 10GbE ports. Figure 4. NetApp FAS8020 IOPS and Throughput as Compared to FAS3220

160,000 3,200

140,000 2,800

120,000 2,400

100,000 2,000 MB/Sec

80,000 1,600 IOPS 60,000 1,200

40,000 800

20,000 400

0 0 FAS3220 FAS8020

IOPS Throughput

Next, ESG Lab looked at the performance of the FAS8000 series when presented with a mixed workload as might be seen in a highly virtualized environment, with multiple virtual machines running different applications accessing the FAS cluster simultaneously. The industry-standard Iometer workload generation utility was used to emulate the I/O activity of four common business application workloads:  Exchange 2013: The Iometer utility was used to generate storage traffic modeled after the I/O patterns produced by the Microsoft Jetstress and Loadsim utilities. ESG Lab used Iometer to simulate the activity of typical Microsoft Exchange users as they send and read e-mails, make appointments, and manage to-do lists. The Iometer utility is, however, a more lightweight utility than either Jetstress or Loadsim. Iometer was designed to focus solely on storage performance.  OLTP Database: The Iometer utility was also used to generate response-time sensitive online transaction processing (OLTP) database traffic.  File Services: The Iometer utility was used to generate file server traffic. The I/O definition was composed of random reads and writes of various block sizes. The file server Iometer profile used for this test was originally distributed by Intel, the author of Iometer; Iometer has since become an open source project.2  Backup Reader: The Iometer utility was used to generate simulated restore traffic. The I/O characterization was composed of 100% sequential 64KB I/O, with 100% reads. Each of the four workloads was run in parallel in its own dedicated virtual machine. Data volumes were presented from aggregates on each of the two nodes in the cluster. As seen in Figure 5, performance scaled as additional virtual machines and application workloads were brought online.

2 http://www.sourceforge.net/projects/iometer

Figure 5. FAS8020 Mixed Workload Performance

90,000 80,000 70,000 60,000

50,000 IOPS 40,000 30,000 20,000 10,000 0 4 workers 8 workers 16 workers OLTP Mail File Services Backup

Table 1 lists the detailed data obtained from Iometer during performance testing. It's important to note that overall I/O scaled proportionally according to the number of VMs generating workloads against the cluster. Moving from four to eight workloads in a two node cluster resulted in an increase of nearly 20,000 IOPS, moving from eight to 16 workloads increased overall IOPS by another 30,000. At this point it was noted that the utilization of the nodes in the cluster was rising. Table 1. Mixed Workload Performance Data on the FAS8020 cluster

Backup Average Host Exchange 2013 File Services Workers OLTP IOPS Throughput Response Time Mailboxes IOPS (MB/Sec) (ms) 4 11,987 93,000 6,917 116 3.63 8 17,239 111,000 13,769 186 5.68 16 39,476 163,000 13,677 613 12.91 ESG Lab again compared the performance of the two-node FAS8020 cluster with the two-node FAS3220 cluster tested in 2013. The results are summarized In Figure 6. As observed in the raw IOPS and throughput tests, the two- node FAS8020 cluster was able to drive more than 3x the IOPS and throughout of the two-node FAS3220 cluster when servicing a complex, highly virtualized mixed workload.

Figure 6. FAS8020 Mixed Workload Performance as Compared to FAS3220

90,000 80,000 70,000 60,000 50,000

IOPS 40,000 30,000 20,000 10,000 0 FAS3220 FAS8020 OLTP Mail File Services Backup

Exchange 2013 Performance and Availability The Microsoft Exchange Solution Reviewed Program (ESRP) is designed to facilitate third-party storage testing and solution publishing for Exchange Server. Microsoft Gold Certified or Certified Storage Partners use the ESRP framework to test their storage solutions in the context of a Microsoft Exchange deployment. The programs combine a storage testing harness (Jetstress) with publishing guidelines. Manufacturers use the ESRP framework to test storage solutions and then submit results to Microsoft for review with approved solution results getting posted on the Microsoft Exchange ESRP website.3 ESRP version 4.0 focuses on Exchange 2013 and Microsoft provides specific recommendations regarding Exchange database size, I/O profiles, mailboxes per server, and mailbox size. Table 2 shows the configuration and assumptions used for ESRP testing of a 200,000 mailbox solution using a six- node NetApp FAS8060 storage cluster. Table 2. 200,000 Mailbox Exchange 2013 Mailbox Resiliency Configuration–Six-node FAS8060 Cluster

Number of Exchange Mailboxes Simulated 200,000 Number of Database Availability Groups (DAGs) 3 Number of servers per DAG 16 (8 tested) Number of active mailboxes per server 8,340 Number of databases per host 6 Number of copies per database 2 Number of mailboxes per database 1,390 Simulated profile: I/Os per second per mailbox 0.06 IOPs per mailbox (0.05 IOPs plus 20% headroom) (IOPS include 20% headroom) Database LUN size 1.5TB Log LUN size 20GB Total database size for performance testing 195.4TB % of storage capacity used by Exchange database 90%

3 http://technet.microsoft.com/en-us/exchange/ff182054.aspx

The publishing guidelines require vendors to specify the server, storage, and Exchange configurations tested. In addition to the Exchange Simulation data shown in Table 2, vendors must also provide details of the configuration and Exchange deployment tested. ESRP differs from standard I/O generation benchmarking suites in three important ways. First, it employs the Jetstress utility to simulate Exchange traffic against real Exchange databases, with logging and file attachments. Second, the testing is designed to measure both the performance and the reliability of a given solution. The performance test runs for two hours while the reliability test runs for 24 hours, and both tests must run without exceeding a prescribed disk latency threshold–average response times of 20 ms or lower for database reads and 10 ms or lower for log writes are required. Microsoft made significant changes to Exchange starting with Exchange 2010 which greatly reduced disk I/O requirements compared to previous versions. This not only increases the number of mailboxes that can be supported by each disk in a solution, but also enables the option to use larger format disks to support larger mailboxes. With Exchange 2010, Microsoft introduced the Mailbox Resiliency feature which maintains redundant copies of message store databases using Database Availability Groups (DAGs). NetApp’s advanced storage virtualization capabilities and integration with Microsoft Windows and Exchange reduce administrative overhead by eliminating the need to pre-provision based on future capacity or performance requirements, and by enabling storage provisioning and high availability configuration directly from the Microsoft Management Console (MMC). As is seen in this analysis, Exchange 2013’s reduced I/O requirements enable users to support more mailboxes in a given system. The goal of ESRP is to verify that a vendor’s storage solution can reliably sustain an Exchange I/O load in a specific configuration with predictable response times. Microsoft makes it quite clear that these tests should not be used for performance comparisons, but in practice, end-users routinely look to these detailed test results to gauge how well a given configuration will perform under specific conditions. This is not a bad use of these results, provided that all relevant factors are taken into consideration.

What the Numbers Mean When considering both NetApp-submitted ESRP 4.0 results and results obtained in the lab using the Iometer workload generation tool, it’s important to note a few things:  NetApp FAS has demonstrated the performance and reliability required to support mission-critical Microsoft Exchange deployments using the ESRP framework.  ESG Lab testing was able to confirm support for nearly as many mailboxes as the ESRP configuration, using a smaller system with far fewer disk drives.  No environment is completely homogeneous and every Exchange implementation will have a mix of different user profiles in unique proportions, so the supportable number of mailboxes will never fall exactly into any one of these scenarios.  Storage is only one factor to be considered when designing an Exchange solution. Users also need to consider server CPU and memory, network infrastructure and latency, recovery requirements, and resources consumed by other applications as well as the specific mix of mailbox sizes and usage patterns that will need to be supported. Microsoft provides extensive information on analyzing Exchange Server performance looking at all of these factors, which are outside of the scope of the ESRP-Storage program.4

4 http://technet.microsoft.com/en-us/exchange/fp179701

Why This Matters Storage scalability and performance are significant challenges for the growing number of organizations embracing server virtualization technology in support of an IT-as-a-service, on-demand delivery model. Storage benchmarks have historically focused on one type of workload (e.g., database or e-mail) and one key performance metric (e.g., IOPS, response time, or throughput). Server benchmarks have typically tested just one server running a CPU- intensive workload that doesn’t stress storage. To help IT managers understand how the NetApp FAS8000 series performs in a highly virtualized environment, ESG Lab ran a benchmark designed to assess how real-world applications behave when running on multiple virtualized servers sharing a single storage system. ESG Lab simulated the traffic of a mid-sized organization running multiple Exchange 2013 servers, several busy OLTP databases, a file server supporting hundreds to thousands of users, and backup to disk, all running simultaneously on just two physical servers and a two-node FAS8020 cluster running Data ONTAP 8.2.1. Performance of the FAS8020 cluster scaled smoothly as additional virtual machines and workloads were brought online and demonstrated excellent host response time throughout the tests, indicating significant available headroom to grow as performance requirements increase. The FAS8020 cluster demonstrated 2-3X performance improvement over the previous generation FAS3220 cluster in every test case, from raw IOPS and throughput to more real-world mixed application workloads. Through careful examination of ESRP results combined with hands-on testing, ESG Lab has verified that the NetApp FAS8000 series can be deployed to provide cost-effective, easy-to-configure storage for Exchange environments of all sizes with excellent scale-out capacity and performance. NetApp Flash Cache enabled a high performance Exchange environment to run on a small number of high capacity drives with excellent response times.

Non-disruptive Availability and Maintenance ESG Lab validated the non-disruptive capabilities of FAS storage running clustered Data ONTAP 8.2 using a cluster of FAS controllers. Clustered ONTAP 8.2 is designed to enable users to manage, upgrade, and service their storage infrastructures without disruption over the life of their data. Clustered ONTAP 8.2 nodes are configured in fault tolerant pairs, with each node in a pair providing failover services for its partner. All nodes in the cluster are managed as a single system.

ESG Lab Testing As seen in Figure 7, ESG Lab began with a cluster consisting of two FAS nodes. A mixed application workload was run during the expansion of the cluster from two to four nodes. Figure 7. Mixed Workload Performance During Upgrade

The cluster was first scaled to three nodes, and then to four. I/O never stopped, and response time stayed low throughout the process. Figure 8. The Data ONTAP 8.2 Cluster, Expanded to Four Nodes

After both new nodes were added–and without interruption to I/O–ESG Lab added additional virtual machines. Each of the four workloads was run in parallel in its own dedicated virtual machine. One instance of each workload was set to run against each of the four nodes in the cluster. Response time stayed low even with 16 VMs running different applications with different I/O patterns.

Next, ESG Lab used the four-node NetApp clustered ONTAP 8.2 system illustrated in Figure 9, and pulled a network cable while a virtual machine running Windows 2008 Server used the Iometer utility to generate a SQL Server workload continuously against a volume on an aggregate in the cluster. The virtual machine was running Microsoft MPIO for path failover. Figure 9. The MPIO Path Failover Test Bed

As seen in Figure 10, the virtual machine accessed the volume over two virtual NICS, each of which was mapped to the clustered ONTAP 8.2 system over a different physical NIC in the vCenter server. Each path was passing data at just under 25% utilization. Figure 10. Path Failover and Failback with MPIO

When ESG Lab disconnected one 10GbE cable that was connected to the server, the traffic immediately dropped to zero. Local Area Connection 3 then resumed I/O, and the utilization increased to nearly 50%, as the single connection took on all the workload. Next, ESG Lab reconnected the cable and watched as Local Area Connection 2 came back online and traffic rebalanced across both adapters.

With the workload still running, ESG Lab moved the 448GB volume to another aggregate, under a different node in the cluster. As seen in Figure 11, this is a simple, GUI driven activity and all that is required is the source and destination volumes. Figure 11. Moving a Volume Under Load

The move completed in less than ten minutes, with no interruption to I/O. Figure 12. Volume Move Completed

Why This Matters Virtualization of servers and business applications increases both data storage requirements and complexity as IT strives to bring applications and services to traditional IT infrastructure dynamically and on-demand. IT administrators and managers were asked by ESG to list their top IT priorities for 2014-15 and increased use of server virtualization was once again a top response, while managing data growth and business continuity/disaster recovery also appeared in the top ten responses in the same survey.5 52% of organizations reported that spending on storage infrastructure and virtualization would increase in 2014. This suggests that data migrations and sensitivity to data availability will increase as storage and servers are consolidated into federated pools of IT resources. As storage environments grow in size and complexity, so does the impact of data outages. Regardless of the number and types of hardware failures that may occur during the life of data on disk, managers, employees, and customers expect their data to be available. NetApp FAS storage systems are designed to provide flexible, tier-one, federated storage, engineered for efficiency, performance, and multi-tenancy in support of private and public cloud delivery of storage services. ESG Lab non- disruptively upgraded a FAS cluster running Data ONTAP 8.2 from two nodes to four while a mixed application workload was running. The FAS cluster scaled non-disruptively as additional virtual machines and workloads were brought online and demonstrated excellent host response time throughout the tests, ESG Lab found the NetApp clustered Data ONTAP 8.2 system able to non-disruptively fail network connections over while a server was executing live read and write I/O to the volume. ESG Lab was able to move a volume to a different aggregate and fail back on-demand, with no interruption to data access and no impact on performance. ESG Lab has confirmed that the FAS series, combined with Data ONTAP 8.2, can provide an always-on, non- disruptive storage environment able to operate through planned maintenance and unplanned faults thanks to a tightly integrated, highly available architecture combined with robust cluster services.

5 Source: ESG Research Report, IT Spending Intentions Survey, February 2014.

Validating Cost-effective, Proven Storage Flexibility ESG Lab examined the NetApp FAS8000 series’ flexibility as a platform, validating the ability of NetApp FlexArray storage virtualization to enable organizations to leverage existing and new heterogeneous storage hardware systems. FlexArray builds on NetApp’s history of storage virtualization features, including NetApp FlexVol storage- virtualization technology which enables organizations to respond to changing storage needs, reduce overhead, avoid capital expenses, and reduce disruption and risk. FlexVol technology aggregates physical storage in virtual storage pools, so IT can create and resize virtual volumes as application needs change. Both FlexArray and FlexVol work in conjunction with NetApp UTA2 target adapters to provide additional hardware flexibility.

ESG Lab Testing ESG Lab began with a look at multi-vendor heterogeneous storage provisioning with NetApp FlexArray. Some storage system vendors are able to integrate their own catalog of disparate storage architectures to present a unified view of the storage system to the administrator and end-user. Unlike these closed environments which only work with that vendor’s specific products, NetApp takes an open approach in order to provide maximum end-user and administrator flexibility. Previously, the NetApp V-Series Open Storage Controller virtualized third party storage solutions. This functionality is now directly integrated into the FAS8000 series as FlexArray. Enabled with a simple license key, IT organizations can now enjoy the benefits of FlexArray third party storage virtualization directly integrated into their primary storage infrastructure. FlexArray provides administrators the flexibility to leverage and pool existing and new heterogeneous storage systems. FlexArray brings the flexibility of presenting a unified storage environment to the end-user while simultaneously enabling Data ONTAP storage functionality on these disparate storage systems (Figure 13). In addition, administrators can manage the entire heterogeneous storage environment using OnCommand, NetApp’s storage management console. Figure 13. FAS8000, FlexArray, and Unified Target Adapter

ESG Lab first started NetApp OnCommand system management software. Within OnCommand is an option to review and install licenses to enable various purchased features. With a few mouse clicks, ESG Lab installed the FlexArray license, shown in Figure 14 as “Virtual Attached Storage License.” Figure 14. FlexArray Licensing

Once installed, FlexArray enabled the FAS8000 series cluster to manage and export storage from a variety of NetApp and third-party storage solutions including NetApp FAS Series, NetApp E-Series, EMC, HP and Hitachi. Next, ESG Lab reviewed the configuration of the latest generation of the NetApp unified target adapter, known as UTA2. This advanced adapter implements two SFP+ ports, which can be configured as either two 16G Fibre Channel or two 10GbE/FCoE. This UTA concept is also implemented on half of the embedded controller ports in each controller. In that case, port pairs on each ASIC are field-selectable for either 16GFC or 10GbE. Customers can change the personality of UTA2 ports in software without having to swap out adapters. Using OnCommand, ESG Lab selected the Ports/Adapters management pane, and then reviewed the Ethernet Ports (Figure 15). The UTA2 installed in the test bench FAS8020 was configured for Ethernet, connected to a physical switch, and appears as Ethernet ports e0c and e0d. The same 2 ports on the same UTA2 also appear as FC/FCoE ports 0c and 0d. Since the UTA2 was configured for Ethernet, the FC/FCoE tab shows ports 0c and 0d as unconnected. With a few mouse clicks, administrators can quickly and easily change the personality of the UTA2 from Ethernet to FCoE. It is important to note that when the personality is set to FCoE, regular Ethernet protocols are also supported and can be mixed with FCoE on the same port.

Figure 15. UTA2 Configuration

Using the NetApp UTA2, administrators can support all block and file protocols using onboard ports in the FAS8000, configured exactly as needed. This simplifies planning and management because each FAS system can be precisely configured to match each organization’s physical plant. Next, ESG Lab reviewed NetApp FlexVol storage-virtualization technology, which enables organizations to respond to changing storage needs, reduce overhead, avoid capital expenses, and reduce disruption and risk. FlexVol technology aggregates physical storage in virtual storage pools, so IT can create and resize virtual volumes as application needs change. NetApp Data ONTAP is different from traditional storage systems in that storage virtualization can be applied at the volume level, commonly used in NAS configurations, or at the LUN level, useful for SAN configurations where not every LUN may be a good candidate for thin provisioning. Figure 16 compares traditional provisioning with NetApp FlexVol thin provisioning. When provisioning a volume for a host in the traditional manner, the amount of physical disk allocated is equal to the size of the volume presented to the host. When thin provisioning by volume, the entire volume is thinly provisioned, providing storage efficiency to any client or server that attaches to the volume. Thin provisioning by LUN provides greater flexibility, enabling administrators to selectively apply thin provisioning only where it makes sense.

Figure 16. Traditional Provisioning Compared to NetApp Thin Provisioning

In the example shown above, a server running a mission-critical online application has a storage capacity requirement of only 10GB, but over time, its storage requirement is projected to grow to 1TB. With traditional provisioning, 1TB is pre-allocated to the application. This means that 99% of the physical storage assigned to this volume would be unused and, importantly, unavailable to other applications. Using NetApp FlexVol technology, an administrator need only specify a volume's virtual capacity. This is the maximum capacity that the volume may consume on disk. Physical disk space is consumed only as data is written to the thin provisioned volume. To test thin provisioning using FlexVol, ESG Lab created a new 3TB volume on an empty aggregate. Inside this volume, a 2TB thinly provisioned LUN was created and assigned to a virtual machine. From the console of the virtual machine, the volume was mounted and initialized using Windows Disk Manager. ESG Lab copied a 12.6GB file to the LUN, and confirmed available capacity by viewing the properties of the disk in the virtual machine, as shown in Figure 17. Figure 17. Capacity Consumed on the LUN

Finally, ESG Lab confirmed that the 2TB volume was only consuming .01TB on the aggregate using the NetApp OnCommand management application, seen in Figure 18. Figure 18. Space Consumed By the FlexVol

Why This Matters ESG has found that end-users often acquire and implement new storage systems when they have allocated but unused storage capacity. In a survey of enterprise storage administrators, more than half reported that up to 50% of their purchased storage capacity was stranded and unused. In addition, 45% percent indicated that they purchased new storage systems to support new and existing applications every six months or more frequently.6 With storage virtualization and thin provisioning, less physical storage is required since the amount of stranded storage is significantly reduced. In addition, companies are continuously challenged to cost-effectively meet the capacity and performance requirements of applications. Failure to meet these requirements can result in downtime leading to lost productivity and costly loss of services. FlexArray storage virtualization, integrated directly into the FAS8000 series and NetApp Data ONTAP 8.2, is able to pool and share third-party storage hardware solutions, providing flexibility and capturing potentially stranded storage. Flexibility is also enhanced with the NetApp UTA2 adapter which is quickly and easily configured for either Ethernet/FCoE or native 16G Fibre Channel, enabling rapid reconfiguration to meet changing demands in the data center without needing downtime to change or add hardware adapters. ESG Lab has also validated that thin provisioning on the FAS8000 series with NetApp FlexVol is easy to manage and extremely capacity efficient. Provisioning additional storage capacity happens entirely behind the scenes while the application remains unaware of any changes. The storage system automatically provisions additional capacity as needed without any manual intervention. The FAS8000 series leverages NetApp FlexVol to provide enterprise class storage efficiency to mid-sized organizations.

6 Source: ESG Research Report, Scale-out Storage Market Trends, December 2010

© 2014 by The Enterprise Strategy Group, Inc. All Rights Reserved. Lab Validation: NetApp FAS8000 Series 21 ESG Lab Validation Highlights  Scalability of performance and capacity with the FAS8000 series was seamless. ESG Lab was able to expand a cluster while running a production workload simulation with no interruption to I/O.  Testing with a variety of business-critical, latency-sensitive application workloads yielded high levels of performance and throughput that scaled predictably as virtual machines were added. Testing with a mixed workload methodology scaled nearly linearly as workloads were added, while Flash Cache ensured that response times measured at the hosts stayed in the single digits throughout testing.  ESG Lab has verified that the NetApp FAS8000 series can be deployed to provide cost-effective, easy-to- configure storage for Exchange environments of all sizes with excellent scalability and performance.  The FAS8020 cluster demonstrated 2-3X performance improvement over the previous generation FAS3220 cluster in every test case, from raw IOPS and throughput to more real-world mixed application workloads.  MPIO integration enabled a virtual machine to continue operations during a path failure with a minimal pause in I/O. Failback was seamless and fast as well, rebalancing I/O across both paths the moment the failed path was brought back online.  The clustered Data ONTAP 8.2 system was able to move volumes between nodes quickly and completely without disruption.  FlexArray Storage Virtualization was directly integrated into FAS8000, and was able to pool storage from heterogeneous third-party storage platforms.  The UTA2 two-port target adapter was easily reconfigured from Ethernet personality to FCoE personality, providing flexible host interfaces without having to change or add hardware adapters.  ESG Lab validated that thin provisioning with NetApp FlexVol is easy to manage and extremely capacity efficient.

Issues to Consider  Adding or removing cluster nodes is a manual process and best practices recommend that it only be performed by skilled and trained personnel. To be fair, adding or removing cluster nodes is not something users are likely to undertake on a regular basis, but some automation of these procedures (e.g., error checking, load balancing, etc.) would be a useful enhancement.  ESG Lab has tested linear scaling with multiple iterations of clustered Data ONTAP, and have repeatedly validated the near-linear scaling of the performance of Data ONTAP and the FAS platform. Thus, for this lab review, ESG Lab chose not to retest linear scaling.

© 2014 by The Enterprise Strategy Group, Inc. All Rights Reserved. Lab Validation: NetApp FAS8000 Series 22 The Bigger Truth ESG research reveals that increased use of server virtualization, improved backup and recovery, and data growth management are all frequently cited important IT priorities.7 The amount and variety of data that businesses need to store is skyrocketing, driving the growth in overall storage use and costs. To address data growth without interrupting business operations, rapid deployment of storage and IT resources in order to meet increasing demand becomes a function of scalability. As IT continues to implement advanced capabilities and services such as virtualization, storage infrastructure becomes more complex. This complexity is multiplied by the rapid growth in data being managed, and increases the demands on both the primary storage and the data protection and disaster recovery infrastructures. As a result, IT is feeling more pressure to provide advanced solutions, and simple, scale-up storage systems are no longer up to the task. Using NetApp FAS8000 Series storage with clustered Data ONTAP, organizations can manage all the heterogeneous storage in their environments as a single pool using a single interface with common tools. Every NetApp FAS system—whether primary or deep archive—runs Data ONTAP. Data ONTAP provides a consistent user interface and powerful storage efficiency technology. All storage can be available to applications and users over either SAN or NAS protocols. NetApp designed the FAS8000 series and Data ONTAP for scalability, including the ability to dynamically assign, promote, and retire storage, replace or upgrade disk shelves, and move data between storage controllers and tiers of storage, all without disrupting users and applications. This enables administrators to increase capacity while balancing workloads, and can reduce or eliminate storage I/O hot spots without the need to remount shares, modify client settings, or stop running applications. FAS8000 supports all Data ONTAP features including Data ONTAP Edge, Snapshot copies, SnapMirror replication, deduplication, FlexClone, and SnapVault, creating an integrated data protection environment that enables backup administrators to protect and restore data and applications quickly and easily, from disk, leveraging automated policies. FAS8000 integrates the Unified Target Adapter (UTA2), supporting heterogeneous high-speed connections to storage clients. With field-selectable native 16GFC and 10GbE, UTA2 enables administrators to scale their environment to meet the ever-changing needs of the growing data center and data storage architecture without the risk of running out of specific port types or needing to pre-specify configurations. ESG Lab recommends a serious look at the benefits that can be realized from virtualizing storage environments with NetApp FAS8000 series running clustered Data ONTAP. NetApp continues to execute on its vision of a single, unified, scale-out platform, delivering seamless scalability in performance and capacity, non-disruptive operations, and field-proven storage and operational efficiency. Through hands-on testing, ESG Lab has confirmed that NetApp can bring a flexible and efficient service-oriented model to heterogeneous storage environments while reducing complexity and delivering a robust infrastructure foundation for shared, on-demand IT services.

7 Source: ESG Research Report, IT Spending Intentions Survey, February 2014.

Storage Each node: 24GB physical memory 1TB VST Flash NetApp FAS8020 (2 nodes) 18 x 1 TB SATA data drives–servicing workload 2 x UTA2 ports, 2x10GbE ports Data ONTAP 8.2.1 RC2 Each node: 12GB physical memory NetApp FAS3220AE (2-4 nodes) 512GB VST Flash 12 x 1 TB SATA drives 4 x 10GbE ports Servers Virtualization Servers (2) vSphere 5.1 Virtual Machines (8) Windows Server 2008 R2 Software and Tools Data ONTAP Version 8.2, 8.2.1 OnCommand System Manager Version 2.0 Iometer Workloads: Exchange 2013 DB: 32KB, 73% read, 100% Random Version 2008.06.18 8KB OLTP, 8KB, 67% Read, 100% Random Backup Reader: 64KB, 100% Read, 100% Sequential File Services: Mixed Block Sizes, 80% Read, 100% Random

20 Asylum Street | Milford, MA 01757 | Tel: 508.482.0188 Fax: 508.482.0218 | www.esg-global.com