NetApp - Engineering Applications

Managing remote office or top flight engineers home office storage?

Earlier this month NetApp released it's pricing and programs around the FAS2000 storage platform.  In addition to new pricing, there was also a new entry level platform released the FAS2040.

What this means for engineering applications managers and director's of programs overseeing them and their remote offices is price sensitive storage solutions designed to support engineering application infrastructures.  We've heard many times over the our best customers in multiple sectors have key employees who work in remote offices or have dedicated home offices.  In these cases they want all the efficiencies provided by NetApp's FAST, zero-tiering, storage efficiency, database and file ready FAS6000 line, but not at that price point or overall footprint.  That's where the FAS2000 line comes in, but more importantly buying the FAS2000 can start as little as $8000 for those small offices and home office users scattered around the world.

From our recent press we shared this information:

Starting at $7,920, the updated FAS2020 offers smaller, price-conscious IT organizations a simple, well-integrated solution for enterprise-class functionality at an unprecedented price point. The FAS2020 now also offers multiprotocol support, including CIFS and NFS, for added flexibility at no extra charge. The enhancements make the FAS2020 one of the industry's best values for the Midsize Enterprise market -- and make simple and flexible unified storage a reality for a broader set of customers.

The best part is all of our other storage efficiency for databases and engineering applications (like perforce software development) capabilities can follow with this unit since we are the industries best in class unified storage.  While the FAS2000 isn't as FAST as the FAS6000; it certainly makes the grade of A+ for your remote or home office situations.  Remember, NetApp's FAST PAM technology is designed to help customers get better value out of their storage, mixing enterprise flash and huge SATA drives, as well as their 15K and 10K rpm media. Our FAST PAM does this by automating the allocation and utilization of available resources.  NetApp FAST PAM delivers engineering results to market faster AND price sensitive solutions!

More training on the FAS2000 software packs

Austin Texas - Tech ONTAP - LIVE!

Join Us for the Austin Tech OnTap Live
NetApp User Group
October 1, 2009

Register Now

Come mingle, enjoy libations and munchies and learn how comprehensive data management just got simpler. CommVault and NetApp continue to demonstrate innovation and market leadership together, as evidenced by the hundreds of joint customers that enjoy the fully optimized disk-based data and information management solutions. Find out how you can better protect, manage and recover critical data anytime and anywhere.

I'm not presenting, but look forward to seeing you there if you can make it!

Cloud Strategies for Engineering Applications

Today NetApp announced its broad plan for Cloud Computing.  The cloud computing launching comprises three core definitions and a fourth core definition that includes four pillars of services.  In all, seven items set the foundation for how NetApp is approaching cloud computing.

• Cloud is defined as “Generally IT as a Service (ITaaS)
• Cloud computing is defined as A business model for delivering IT as a Service
• Types of clouds are defined as a) Private Cloud being enterprise internal and external b) Public Cloud being Non-IT or IT only

Variations of cloud services, remember “cloud” is “Generally ITaaS.”  The variations serve as the four core pillars of cloud services.

• Infrastructure as a Service
• Platform as a Service
• Software as a Service
• Storage as a Service

 

To complement this strategy announcement NetApp coupled with this several other key announcements around supporting technologies.  These include Data ONTAP 8, which includes support for very large data containers, which we call 64-bit aggregates.  Think of these as unending pools of storage for your highest engineering applications growth.  For example, if required, you could support an NFS volume that encompasses 100TBs of storage.  Moreover, Data ONTAP 8 includes key services for nondisruptive data mobility, ensuring that large amounts of data can be managed without impacting large numbers of users.

While Data ONTAP 8 becomes available in the September, so do two other core technologies for engineering applications and infrastructures wanting to take advantage of the cloud options available from NetApp.  These two technologies are a large capacity flash read-cache module for FAS systems called PAMII and a new high-density disk shelf from NetApp called the DS4243.

DS4243 - Higher Density SATA Shelf (versus DS14)

 

PAMII Flash Read Cache Card

Further, coming in early 2010, NetApp will be delivering Data Motion ™.  Data Motion is NetApp’s new data mobility technology enabling enterprises (and service providers :-) ) to move data nondisruptively across storage systems, think of this as platform tiering vs disk tiering (there’s really no need to tier data from disk type to disk type when using PAMII).  Data Motion is built on NetApp’s unified architecture and leverages the built in storage virtualization and multi-tenancy capabilities of Data ONTAP.

Now that we have set the baseline, defined cloud, the four core pillars of cloud services and some technology used to enable cloud computing, how do you make it all work in Engineering Applications?

First and fore most, reduce your dependency on more disk spindles by architecting your engineering application infrastructures with PAMII.  PAMII enables you to improve IOPS and reduce latency for random read intensive workloads.  Moreover, PAMII uses three caching modes and can be sized up to 512GB, so the application for PAMII is quite broad.  Key benefits of introducing PAMII into your engineering app environment are:

• Optimize performance at a lower cost
• Automatically tiers active data to higher performance storage
• Get the IO throughput, without impacting data center square footage
• Grow capabilities, without impacting cooling output and power consumption
• Great for engineering applications, file services, databases, and virtual infrastructures

 

Here are some key aspects of PAMII to consider when designing or updating your engineering applications infrastructure.

 

While it is easy to toss around the idea of flash cache, what happens when the application starts pushing back on the storage system?  Well we put these systems to the test with SpecSFS 2008 and here are some graphical plots of the results.

 

Well we didn’t want anyone guessing, so we are making it perfectly clear that with PAMII you get higher IOPS with less latency using FEWER FC disks for the nominal SFS workloads.  However, we didn’t stop there, we took the testing to a whole new level – we introduced SATA (remember that high density SATA shelf we just delivered, the DS4243, it applies here).

When comparing to SATA to FC you would expect SATA to be clearly behind. However, with PAMII it’s hard to tell who is coming in first.  However, don’t think this is just a speed test.  Look at the total capacities.  This is the equivalent of a VW minivan keeping up with a Porsche on the race track (http://www.youtube.com/watch?v=SsaBYyjhHuo jump to 2:37 to get the full effect of PAMII with SATA drives).  Granted it’s not sexy, but it’s fast and it can haul more data – I’m talking about SATA and PAMII.  That VW bus video, well try explaining that one to your co-workers...

While testing with SpecSFS2008 is the standard or minimum bar of entry, think of it as the crash test ratings.  If your storage or OS vendor doesn’t take the test, you probably should move on.  Otherwise you’ll be trying to drive with SUN in your eyes.  So, not leaving stones unturned, we tested two key engineering applications – Perforce for Software Configuration Management (SCM) and Siemens Teamcenter for Product Lifecycle Management (PLM).  Rather than try explaining all the nuances of multiple engineering applications, let’s focus on this graphic.

In this graphic we explain, visually, the impact on a single developer’s development cycle.  Technologies like PAMII can improve the cycle time of the sync and integrate processes a developer waits on while developing software or testing software.  Sync is syncing the code and integrate is integrating the code lines together with the main SCM.  Reducing these artificially inflated times by 50% increase the work done across hundreds of developers over several months, resulting in development cycle times that can be reduced by many months over the course of a single project.  If you have multiple projects (as we all do) then you should see reduced cycle times, resulting in faster time to market and marked first mover opportunities across all economies and sectors.

The one topic I didn’t spend much time on today was NetApp Data Motion, but suffice it to say that as you deploy your engineering applications on NFS, iSCSI, etc.  You’ll find that virtualizing storage using NetApp Data Motion and NetApp Multistore, will yield the most positive results.  So as you move forward to architecting your dynamic data center for engineering – apply all the technology I shared today.

Reference:

keyword engapps on netapp.com sites

TR 3718 - Best Practices for Deploying a Perforce SCM System by Using NetApp Storage Systems

TR 3658 - Siemens PLM Software Teamcenter on NetApp Storage over NFS: A Reference Architecture

WP 7061 - Flash Memory Technology in Enterprise Storage – Flexible Choices to Optimize Performance

pNFS is it more than hype?

Hype Cycle for Storage Hardware Technologies, 2009

Looking at Gartner's most recent Hypecycle, pNFS is "On the Rise"

That's all I can say about the Hypecycle, you'll have to read the document yourself.

Learn more about pNFS at SNIA's NFS SIG

The NFS SIG is focused on creating a portfolio of marketing and educational materials explaining the features, benefits, use cases and best practices related to NFS-based NAS solutions.  Particular focus will be on applications and benefits of emerging new capabilities, such as pNFS.  The goal is to drive the continued market growth and broad adoption of NFS, and a rapid market uptake as vendors bring their pNFS implementations to market.

NFSv4.1/pNFS Special Interest Group.

JK

Free SHIP and BEER with NFSv4

Check it out over in our communities site

FREE NFSv4 TRAINING

The primary business benefits for NFSv4 are user experience (performance), intellectual property management (security/internationalization), capital cost reduction (security) and business cadence/operations (high availability).  We have several customers that depend on NFSv4 to fuel there Tibco EMS architecture because Tibco, NetApp's NFSv4 and Data ONTAP 7G support for key elements:

 

1) write order, 2) synchronous write, 3) distributed file locking, 4) unique write ownership (you can read more about this in my blog series below)

 

 

Engineering Applications Technical References (TRs)

It's been a few months since I posted, between being sick and the too busy - phew.  Time is certainly finite, the responsibilities I have are not :-)  I was hoping to write a bit more on these two technical references, but given the constraints of work and balence with life I'm simply sharing them with you.

I'll get back to NFSv4 and the update from SNW Spring 2009 later this month.  On a bright spot - the NFS SIG at SNIA is going well.  I just had another SNIA session approved for Fall in Phoenix, AZ!  The title of the talk is pNFS, parallel storage for grid, virtualization and database computing.  We'll be covering use cases on virtualization and databases - woo hoo!

Best Practices for Perforce - 3718

This white paper provides recommendations around the NetApp storage architecture for hosting Perforce SCM systems based on your organization’s performance, high-availability, and scalability needs. For this purpose, NetApp conducted Perforce benchmarking tests on various storage architectures. This paper discusses the pros and cons of the various storage architectures available for deploying Perforce databases and depot files on NetApp storage and provides deployment guidelines for an optimal distributed development infrastructure, as well as alternative approaches for speeding the creation of user workspaces.

Key Take Aways

• Perforce SCM Architecture
• Advantage of Deploying Perforce on NetApp
• Performance Benchmarks
• Deployment Options
• Data Protection Options
• Best Practices for Deployment of Perforce
• Distributed Development Architecture
• Advantage of Deploying User Workspace on NetApp

Best Practices for Teamcenter - TR 3658

NetApp and Siemens PLM Software conducted a series of performance and scalability benchmarks as a joint effort to illustrate a reference architecture (two-tier) for Teamcenter with NetApp® storage. This report documents performance benchmarks across SAN and NAS with different workloads (low [<100 users] and medium [<1,000) users], along with the performance impact of additional cache memory in the storage controller on read workloads. The intent of this report is to help organizations choose the optimal storage deployment option, based on performance and total cost of ownership. This document discusses the impact of using NetApp storage for Teamcenter files (tcengvault) and Oracle® Database. The scalability test was run on Teamcenter 2005 SR1. Due to a bug in the Teamcenter 2005 SR1 backup_mode utility, further tests were run for backup and restore, PAM, and deduplication using the Teamcenter 2007 MP3 release. However, the scalability and performance testing with and without PAM apply to both the 2005 SR1 and 2007 MP3 releases. Note that the backup_mode utility can be used only with 2007 MP3.

Key Takeaways

• Teamcenter 2005 SR1/2007 MP3 Architecture
• System-level Scalability Benchmarks (SAN/NFS)
• Read Benchmarks with and without PAM
• Deduplication Testing
• Data Protection Options
• Multi-site Deployment

SNW Recap – Parallel NFS (NFSv4.1)

April was an exciting month for Parallel NFS (also known as NFSv4.1).  On the 6th of April a Parallel NFS presentation was delivered at Storage Networking World in Orlando, FL.  The session was attended by about thirty (30) people.  Many of the attendee’s were ready with pen in hand and paper for writing – as I began the talk it was encouraging to see most people taking notes.

The agenda is far reaching in that we cover the following key topics:

• Introduction to NFS
• NFSv4 – Security, High-availability, Internationalization and Performance (follow the series Pt1,2,3,4,5)
• NFSv4.1 – Status and Overview
• pNFS – Layout Overview; files, blocks and objects
• pNFS – Open Source Client Status

In the introduction to NFS I cover the basics.  Basics include NFS history, server and client availability.  NFS is available in all common operating systems, including hypervisors, such as VMWare ESX Server.  NFS is designed to make data (storage) on file servers available to any NFS client on the network.  NFS Servers were the inspiration to NAS and appliance used today.  Our focus was to cover the basics of what is pNFS, similar to our TechOn Tap article.

 

Parallel NFS is part of NFSv4.1. NFSv4.1 is a minor version upgrade to NFSv4.  In some cases Parallel NFS is referred to independently of NFSv4.1, but pNFS is inclusive of NFSv4.1.  To that end, NFSv4.1 is both a file protocol and storage layout protocol.

 

For traditional NFS client and server access NFSv4.1 will add sessions, directory delegations and improved security with respect to access control lists.  Existing block and object storage systems can leverage NFSv4.1 to read storage layouts for striped LUNS or object targets – “pNFS.”

 

Directory delegations enable an NFSv4.1 client to manage read and writes for an entire directory locally flushing updates back to the NFSv4.1 server.  Other benefits of delegations include metadata access for files and directories can be managed locally.  However, delegations can be recalled by the server if another NFSv4.1 client writes the directory or file.  Design architectures sensitive to delegations can prevent recalls from affecting performance.  Sessions enable NFSv4.1 clients to open multiple NFS file read-write sessions with a single NFS server, i.e. a 10GbE network link could be fully saturated with multiple NFSv4.1 client-server sessions.  Thus delegations combined with sessions will improve performance in NFSv4.1, testing forthcoming.

 

NFS (v3, v4 and v4.1) is a great replacement for LUN based cluster file systems based on cost, managementand client density.  The latter, client density, is one of the key use cases for parallel NFS.  Specifically, a use case was something that we didn’t include in the session on Parallel NFS at SNW.  However, client density is important across multiple environments – in decision support systems, enterprise collaboration, home directories, software development and high-performance computing.

 

In the session I focused on pNFS ability to support other storage protocols by managing the layout of the storage in a high-performance stripe.  One of the audience questions was where pNFS is concerned, what workload or business application is relevant?

 

Next up is a use case for pNFS around virtualiztion and pNFS panel where we mentioned the SNIA NFS Special Interest Group (SIG)...

 

Disambiguate your deduplication with primary system rehydration for engineering applications

Unstructured file data is growing faster than database storage and makes up 50% or more of all enterprise storage. In the April issue of Tech OnTap, we described how you can substantially cut your storage requirement by consolidating and deduplicating your unstructured data—including engineering data. It turns out that deduplication applied to applications such as Siemens Teamcenter can recover 57% or more storage space. One company saw savings of 40% to 70% deduplicating volumes storing CAD, Office and other document types.

Check out Dedupe Unstructured Data for Up to 70% space savings

Are you being overwhelmed by file data? What do you think about the advantages of this approach? Does it make sense in your environment? Why or why not?

Using primary system deduplication is critical in an enterprise environment.  If you consider or use inline deduplication that's not primary system integrated, like Data ONTAP deduplication, can leave your data inaccessible.  Deduplication is like encryption if the deduplication systems are removed or become unavailable. Don't get duped by the competition, use NetApp FAS and vSeries to deliver deduplication with a focus on performance and reliable data hydration.

  Inline vendors may lockup your data...

 _{Photo by: Joshua Konkle, SNW Spring 2009}

Finally, read how deduplication can be extended from the storage system, clustered or not, to the client using extension to NFS that Mike Eisler has proposed Storage De-duplication awareness in NFS.  Using NetApp, now you can deliver data efficiency using NetApp Data ONTAP deduplication; in the future extend it to the desktop with aware file protocols, like NFS.

Landmark PetroStor; Texas-Tea, Black Gold...

I live in Austin, TX so I'm always looking for opportunities to call Oil Texas Tea.  Plus, for those that know the reference about a bubbling crude, you can get that theme song stuck in your head for the rest of the day.

Anyway, I'm blogging because I just found out about this awesome storage solution NetApp built with Halliburton's Landmark.  Landmark provides software to interpret layers of rock, oil, gas and other geophysical conditions related to raw energy resource discovery.

Halliburton’s Landmark announced the availability of its PetroStor™ scalable disk storage platform, which provides customers with petabytes of online capacity. Tested and qualified with Landmark’s exploration and production (E&P) software, the solution is ideal for companies that need fast access to seismic files and archived project data.

Unlike traditional tape alternatives, the PetroStor storage solution provides upstream users instant access to all their seismic files and project data archives. Combining enterprise-class storage from NetApp with real-time compression from Storwize, the PetroStor storage solution lowers storage costs to less than $1,000 per terabyte. It enables storage administrators to increase their storage capacity, along with data reliability and performance, while reducing associated infrastructure footprint, cooling and power costs, as well as administrative costs associated with tape transcription and re-mastering.

The PetroStor storage solution is offered and supported by Landmark, which also offers comprehensive managed services including deployment, data migration and disaster recovery. Landmark can help users migrate their archives from tape to disk and also index files according to their specific architecture, as well as incorporate metadata that makes users’ files easy to find in their ever-growing libraries of data.

Read more about it here in February's installment of Zeus Technology Magazine.

IBM Rational ClearCase can use NFSv4

I was doing some research on NFSv4 the other day and spotted this publication with IBM that shows IBM Rational can and will support NFSv4.

"...NFS V4, only AUTH_SYS authentication is supported for use with Rational® ClearCase®"

The net-net is that for companies wanting to leverage several products offered by NetApp with IBM Rational, they can do so while taking advantage of Unified ACLs between NFSv4 and CIFS and file delegations for software builds.  Moreover, NFSv4 offers a unique feature called NVERIFY which reduces the operations required to perform getattr requests.

Organizations using Rational with 3rd party storage can leverage NetApp's intellectual property around primary storage deduplication as well as the ground breaking performance available in the FAS3100 series with Performance Acceleration Module.  Finally, combine this with a recent offering to guarantee space savings on competitor storage when using Virtualization.

So, If you are software development firm using IBM Rational and interested in using new features in NFSv4, NetApp has the only system that delivers on three critical factors required for success in turbulent economic times:

• Performance - FAS3100 + Performance Acceleration Module (read caching for builds)
• Management - Deduplication for greater density in source code and operating system virtualization
• Reliability - RAID-DP and Snapshots are NetApp originals which keep your running

Ensure you engineering business unit can get control, delivery consistency to offer a development cadence that delivers within your trade-off triangle - cost, people and schedule.

Later this month you'll be hearing about some new engineering partnerships that should encourage you to consider NetApp software and storage for your engineering application platform with Siemens Teamcenter and PTC product lifecycle management solutions.

IBM Rational References:

• FlexVol and FlexClone in IBM Rational ClearCase Environments
• Data Protection: IBM Rational ClearCase Environments
• NetApp NAS solves multi-protocol support for IBM