Transcript
Lab Validation Report Avere FXT Series Efficient Scale-out NAS
By Vinny Choinski
August 2011
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
2
Contents Introduction .................................................................................................................................................. 3 Background ............................................................................................................................................................... 3 Avere FXT Series ....................................................................................................................................................... 4
ESG Lab Validation ........................................................................................................................................ 5 Getting Started ......................................................................................................................................................... 6 FXT Series Scalability................................................................................................................................................. 8 FXT Series Performance .......................................................................................................................................... 10
ESG Lab Validation Highlights ..................................................................................................................... 14 Issues to Consider ....................................................................................................................................... 14 The Bigger Truth ......................................................................................................................................... 15 Appendix ..................................................................................................................................................... 16
ESG Lab Reports The goal of ESG Lab reports is to educate IT professionals about emerging technologies and products in the storage, data management and information security industries. 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 Avere Systems.
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.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
3
Introduction This report documents the results of ESG Lab hands-on testing of the Avere FXT Series NAS appliances. The Avere FXT Series appliances are designed to provide efficient high performance data tiering and scale-out clustering to any NAS solution by separating performance scaling from capacity scaling to deliver each more efficiently.
Background NAS data continues to see fast-paced growth while management and efficiency problems continue to plague storage administrators. Administrators are challenged with managing multiple NAS solutions with a growing number of large file systems. According to ESG research, one-fifth of users are reporting NAS capacity growth of more than 50% per year (up from only 13% who reported greater than 50% annual NAS growth in a 2008 ESG storage survey), and 54% of organizations with NAS installations are indicating that their storage capacity requirements are growing by at least 20% per year. 1 As organizations look for ways to effectively control and manage enormous NAS data growth, business managers are looking to their IT organizations to deliver smart, efficient, and cost effective solutions. As shown in Figure 1, improved business process, better return on investment/speed of payback, and OPEX and CAPEX reduction rank among the top five considerations for justifying IT investments over the next 12 to 18 months. 2 Figure 1. Most Important Considerations for Justifying IT Investments Which of the following considerations do you believe will be most important in justifying IT investments to your organization’s business management team over the next 12-18 months? (Percent of respondents, N=611, three responses accepted) Reduction in operational expenditures (i.e., labor, utilities, materials and other ongoing costs associated with operating an application, system, or business process)
43%
Business process improvement
39%
Return on investment / speed of payback
37%
Improved security / risk management
35%
Reduction in capital expenditures (i.e., purchases of equipment and other fixed assets)
24%
Improved regulatory compliance
19%
Reduced time-to-market for our products or services
16% 0%
10%
20%
30%
40%
50%
Source: Enterprise Strategy Group, 2011.
1 2
Source: ESG Research Report, Scale-out Storage Market Trends, December 2010. Source: ESG Research Report, 2011 IT Spending Intentions Survey, January 2011.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
4
Avere FXT Series The Avere FXT Series is a scale-out NAS platform that provides efficient performance and capacity scaling for NAS environments. As shown in Figure 2, the FXT Series is deployed between the NAS filer(s) and its clients, acting as high performance storage media for active data. Thus, storage capacity can be cost effectively scaled on the backend filers while performance scale is handled on the front end by the FXT Series. All back-end NAS filer storage can be presented and managed centrally via the FXT Series, enabling simplified administration and namespace management. Figure 2. Avere FXT Series
The FXT hardware platform comes in a 2U form factor supporting both 1 GbE and 10 GbE networking. The nodes can be clustered for high availability and are offered in the following configuration options: • • •
FXT 2750 – 512 GB SSD (SLC), 72 GB RAM, 2 GB NVRAM, 2 10GE & 6 1GE ports FXT 2550 – 3.6 TB HDD (SAS), 72 GB RAM, 2 GB NVRAM, 2 10GE & 6 1GE ports FXT 2300 – 1.2 TB HDD (SAS), 64 GB RAM, 1 GB NVRAM, 2 10GE & 2 1GE ports
The FXT Series separates NAS service functions from mass storage functions. It uses a tiered file system (TFS) schema to dynamically place data on the most optimal storage media. Active data is moved to fast media on the FXT Series, enabling dense, cost effective SATA storage to be used on the back end for more persistent data. The cluster can be configured with an unlimited number of nodes for linear performance scale. The FXT Series provides comprehensive monitoring and global namespace management across all nodes.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
5
ESG Lab Validation ESG Lab performed hands-on evaluation and testing of the FXT Series at an Avere Systems facility in Pittsburgh, Pennsylvania. Testing was designed to demonstrate how the FXT Series efficiently provides high performance and capacity scalability for scale-out NAS environments. Also of interest were ease of use and implementation. The configuration used for ESG Lab testing is shown in Figure 3. A three-node Avere FXT 2750 cluster was configured between a pair of NetApp FAS3240 filers and a pair of VMware ESXi hosts each with 24 virtual machines. An Arista DCS-7148S-R network switch was used for LAN connectivity between the host, the FXT Series, and the NetApp filers. Each VMware ESXi host was configured with 12 VMs that accessed data through the FXT 2750 cluster and 12 VMs with direct access to the filers. Across both ESXi hosts there were 24 VMs accessing data through the FXT 2750 cluster and 24 VMs with direct access to the filers. Network VLANs were set up for each configuration. Workloads were installed on each set of 24 VMs to compare performance with and without the FXT Series in the data path. Figure 3. Test Bed
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
6
Getting Started Preparing the FXT Series for use requires performing the initial setup of the cluster from pre-racked nodes with power and network connectivity already provided. The nodes were racked in the lab test bed with VLANs and NAS storage pre-configured. ESG Lab Testing ESG Lab began testing the FXT Series by powering up a node and connecting to the FXT Series setup page via a web browser. The setup process was straightforward and easy to navigate. As shown in Figure 4, the device is configured from a form template within the setup web page—it can be configured manually by filling in the form or by pointing to a configuration file. ESG Lab conducted the manual process using the configuration template. The form required standard network and naming information, with help fields available for most entries. Figure 4. FXT Series Setup Page
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
Next, ESG Lab explored performance-monitoring capabilities by selecting the Dashboard tab from the FXT Series GUI. The GUI allows multiple parameters to be set up and monitored. Figure 5 shows the Cluster Wide – Ops/Second view of the configuration. ESG Lab selected client IOs/sec and MASS storage synchronous and asynchronous IOs/sec to compare the performance seen on the client side with the load on the backend MASS or NAS filer storage. Figure 5. Performance Monitoring
Why This Matters Installing and configuring traditional large-scale file systems can be extremely complex and time-consuming. ESG research found that among organizations considering using scale-out NAS solutions, faster provisioning times and ease of use ranked fourth and fifth among top five drivers for interest in scale-out NAS infrastructure. 3 ESG Lab found that the FXT Series from Avere Systems could be non-disruptively deployed in an existing NAS environment, enabling improved performance and scale. The solution was easily configured from a web-based setup page in a matter of minutes. Ongoing administration proved intuitive and easy to navigate.
3
Source: ESG Research Report, Scale-out Storage Market Trends, December 2010.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
7
Lab Validation: Avere FXT Series
8
FXT Series Scalability Scalability describes the process of non-disruptively adding nodes to a running cluster for performance scale and configuring additional storage capacity on the back end of the solution by adding new exports or NAS filers. ESG Lab Testing ESG Lab validated the FXT Series solution’s ability to deliver high performance while increasing backend capacity for scale-out NAS environments. As shown in the middle of Figure 6, adding cluster nodes to the solution allows performance to scale linearly. The right side of Figure 6 depicts the ability to add capacity with central management via the addition of a NAS filer. Figure 6. FXT Scalability
As shown in Figure 7, ESG Lab used the FXT Series GUI to add a third node to an existing two-node cluster. Node (vantage25) was selected from a list of available nodes. A simple click on the “Allow to join” tab launched the process in the background. Figure 7. Performance Scalability
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
9
Lastly, ESG Lab used the FXT cluster to build a single global namespace that spanned multiple back-end filers. As shown in Figure 8, export (/vol/bryansim) on filer (grapnel) could be found at namespace path (/ESG/HR). Using the Avere GUI, ESG Lab added an existing export (/vol/cust) from filer (northhill) to the same namespace (/ESG) at path (/ESG/finance). This allowed directories from multiple back-end filers to be presented to a host under a common mount point. Figure 8. Capacity Scalability and Simplified Namespace Management
Why This Matters Keeping pace with the runaway capacity demands of growing volumes of data while trying to maintain acceptable performance levels can present major challenges to any It organization. ESG research indicates that 32% of those considering scale-out solutions are looking for improved throughput while 28% see improved IO/sec as factors driving their interest. Additionally, 33% of respondents considering scale-out NAS indicated that they were interested in “ease of management. 4 ESG Lab has confirmed that adding performance and storage capacity to an FXT Series solution is seamless and easy. For performance, a third FXT node was non-disruptively added to a running two-node cluster. After the node was added, the workload was automatically balanced between the three units. Capacity was simply added by presenting a back-end NAS export through the FXT Series. This allowed clients to access exports on separate filers through a single namespace presented by the FXT cluster and enabled simplified and centralized management of multiple backend NAS devices via a single FXT user interface.
4
Source: ESG Research Report, Scale-out Storage Market Trends, December 2010.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
10
FXT Series Performance This section describes the ability of the FXT Series to efficiently deliver reliable performance to scale-out NAS environments. Avere Systems achieves efficient NAS performance by separating performance scaling from capacity scaling. Active data is automatically moved to the FXT Series hardware platform while less active data remains on backend NAS storage. This schema aligns data based on activity and access pattern to the most appropriate storage media. ESG Lab Testing ESG Lab tested IO performance for the FXT Series using the Iometer workload generation tool. Iometer is an opensource storage workload generation tool that is available from www.iometer.org. Iometer was configured to generate a random write and a random read workload for two groups of VMs. The first group was directly connected to a NAS filer pair, the second group with an FXT 2750 cluster configured between the VMs and the NAS. As shown in Figure 9, the FXT outperformed the direct NAS configuration by 9.1 times for the random write workload and by 22 times for the random read workload. A 240 GB data set was used in both the random write and random read tests. Figure 9. More IOs/sec with Avere FXT Application IO 30,000 25,000
IO/sec
20,000 15,000 10,000 5,000 0
Random Write Direct NAS
Random Read Avere FXT
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
11
ESG Lab also tested throughput performance for the direct NAS configuration and the FXT 2750 cluster. An Iometer sequential read workload was used with 24 GB, 120 GB and 240 GB datasets. As shown in Figure 10, the FXT 2750 cluster outperformed the direct NAS configuration by 1.2 times to 3.4 times as the dataset size increased. Figure 10. Throughput vs. Dataset
ESG Lab validated the performance of typical software development tasks by auditing simulated patch and compile process results created by the compilebench utility. As shown in Figure 11, the FXT 2750 cluster outperformed the direct NAS configuration, completing the patch process 4.9 times faster and the compile process 3.2 times faster. Figure 11. Improved Throughput with Avere FXT CompileBench Throughput Results 800 700
MB/sec
600 500 400 300 200 100 0
Patch
Compile Direct NAS
Avere FXT
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
12
Next ESG Lab audited Avere’s published SPECsfs results. As shown in Figure 12, the FXT Series was able to generate 1828 (ops/sec/disk) while still maintaining a good overall throughput of 131,595 (ops/sec). As detailed in Table 1, it should be noted that the FXT configuration was the only configuration in the summary chart to leverage SATA disk and the total number of disks was a fraction of the other configurations tested. Figure 12. SPECsfs Results Summary
SPECsfs Results Summary 2,000 1,800 1,600 1,400 1,200 1,000 800 600 400 200 0 Avere FXT
EMC VNX
NetApp FAS6240
EMC Celerra
NetApp FAS6080
EMC Isilon
SPECsfs (ops/sec/disk)
Table 1. SPECsfs Results Summary Detail
SPEC
Avere FXT
EMC VNX
NetApp FAS6240
EMC Celerra
NetApp FAS6080
EMC Isilon
(ops/sec/disk) File systems Throughput (ops/sec) ORT SSD FC/SAS SATA Total disks
1828 1 131,591 1.38 0 48 24 72
1098 8 497,623 0.96 436 21 0 457
662 2 190,675 1.17 0 288 0 288
446 4 135,521 1.92 0 304 0 304
357 2 120,011 1.95 0 336 0 336
331 1 1,112,705 2.54 140 3220 0 3360
What the Numbers Mean • • • • • • •
On average, the FXT Series achieved a 6X reduction in disk, power, and space requirements. Avere demonstrated linear performance scaling with single-node results of 22k ops/sec, 2-node results of 44k, and 6-node results of 132k. There were no expensive SSD devices used in the FXT solution. The total number of drives for the FXT Series was a fraction of the number of disk used by the other configurations in this SPECsfs summary. The FXT solution leveraged a commodity Supermicro server with a Linux OS and 24 SATA drives as the backend NAS filer. None of the other solution leveraged dense, inexpensive SATA drives. Avere could achieve similar results with a SATA-based backend NAS filer such as NetApp FAS 2000/3000 Series or EMC/Isilon X-Series. The second highest (ops/sec/disk) solution in this summary was configured 436 SSD drives.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
13
Finally, ESG Lab audited real world results from an end-user environment. The client environment compared the performance of an Avere FXT 2700 two-node cluster to an enterprise-class direct NAS configuration with solid state cache and 15k FC drives in an Oracle database environment. Typical daily business processes were run against both configurations on a 1 GB and 10 GB network. As shown in Figure 13, an Oracle bulk update process was run with one through eight clients. The FXT 2700 was able to process 2.8 times more updates at six clients than the direct NAS configuration. Figure 13. Oracle Bulk Update Performance
Why This Matters Maintaining cost effective performance with efficient scalability is a both a priority and challenge for most scaleout NAS environments. Recent ESG research indicates that improved scalability, improved performance, and lower cost of infrastructure rank as the top three drivers for considering implementing scale-out NAS solutions. 5 ESG Lab conducted performance testing simulating typical user and application activity and audited actual end-user results to confirm performance improvements of up to 22X with the FXT Series, depending on the workload. ESG Lab also audited the published SPECsfs results of the FXT Series and validated a 6X average reduction in disk, power, and space requirements as compared to similar solutions.
5
Source: ESG Research Report, Scale-out Storage Market Trends, December 2010.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
14
ESG Lab Validation Highlights ESG Lab confirmed that the FXT Series hardware platform was easy to set up in just minutes. Once racked and supplied with power and network connectivity, the cluster was easily configured from a web-based setup page. The setup page was form-driven with help fields for most entries and required only standard network and back-end NAS access configuration information. ESG Lab validated performance scalability by adding an FXT Series node to a running cluster. The process was seamless and non-disruptive. The process was conducted directly from the user GUI. The join node process automatically matched the firmware of the joining node to that of the running cluster. ESG Lab confirmed that shares and exports from the backend NAS appliance can be managed from the FXT Series GUI. This enabled centralized manage and a simplified namespace for a multiple filer environment. ESG Lab also confirmed that the FXT Series provided efficient performance for VMware environments. Performance improvement was demonstrated to be up to 22x on reads and 9.1x on writes with two Avere nodes, with linear performance scaling available through clustering to go well beyond this. Performance was tested in both the lab test bed environment and through audits of real-world environment test results. ESG Lab was impressed with performance delivery not only for dense, cost efficient back-end SATA, but also for performance-oriented enterprise-class NAS devices.
Issues to Consider The FXT Series has no class of service policy to distinguish between SSD- and SAS-based FXT Series hardware platforms. Avere clusters are deployed as SSD or SAS-only today. The maximum per node SSD configuration is currently 512 GB. There is currently no “progress status bar” available in the FXT Series GUI. It would be good to know that a launched process was still running while conducting administrative tasks. ESG Lab would like to see a built-in data migration option for the FXT Series. Data migration via the FXT Series would add enormous flexibility for NAS replacement/refresh and consolidation.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
15
The Bigger Truth Though typically easy to deploy and provision, rapid growth in a NAS storage solution can quickly introduce management and cost control nightmares. Fifty-four percent of respondents in a recent ESG survey indicated their NAS data was growing at rate of 20% per year, essentially doubling NAS capacity every five years. Combine this with the fact that many NAS solutions are purpose built for either performance or capacity to address specific application challenges, and storage administrators can quickly find themselves managing multiple NAS devices from different vendors each with unique management interfaces. Purpose-built solutions can drive up both OPEX and CAPEX simply by the fact that they typically require multiple technologies and skill sets to address different application requirements. In recent validation testing at Avere Systems, ESG Lab explored the usability and performance attributes of the FXT Series hardware platform. ESG Lab performed typical setup and management tasks and found the FXT series easy to navigate and administer. It can be easily and non-disruptively deployed or removed from an existing and running NAS environment. ESG Lab also confirmed the ability to seamlessly scale the test solution by adding a third FXT 2750 to a running two-node FXT cluster in a test environment. Performance results for the test environment were validated and audited, as were real-world customer experiences. ESG Lab confirmed that by separating NAS performance scaling from capacity scaling, Avere Systems has developed an efficient and cost effective solution for scale-out NAS environments. With the FXT Series, active data is automatically moved to high performance media while larger persistent data sets remain on dense, cost effective SATA storage. This approach is effective for both enterprise class filers as well commodity NAS filer solutions as it eliminates the need for large amounts of expensive SSD devices on a NAS appliance by smartly moving only the active data to the high performance FXT platform. The FXT Series is also non-disruptive to back-end NAS processes such as backup and restore. Existing backup/restore solutions can be leveraged, requiring no FXT Series specific configuration changes. ESG Lab was impressed with the ability of the FXT Series to provide flexible cost effective control of NAS sprawl, enable efficient performance and capacity scale, and centrally manage heterogeneous NAS environments.
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
Lab Validation: Avere FXT Series
16
Appendix Table 2. ESG Lab Avere FXT Test Bed
Avere FXT Appliance
(3) Avere FXT 2750 Appliances
2U Appliance 512 GB SSD 72 GB RAM 2 GB NVRAM 1/10GB Network Interfaces
Storage (2) NetApp FAS
Model FAS3240 Data ONTAP 8.0.1
Network (2) Arista Switch
1U Rack Mount Network switch Model DCS-7148S-R Software Version 4.6.1
Software VMware ESXi Oracle Iometer Compilebench
Version 4.1 Version 11.2.0.2 2008.06.18 Version 0.6
© 2011, Enterprise Strategy Group, Inc. All Rights Reserved.
20 Asylum Street | Milford, MA 01757 | Tel:508.482.0188 Fax: 508.482.0218 | www.enterprisestrategygroup.com
