NetApp Announces Next Gen Hardware – FAS2600, FAS8200, FAS9000 and More

September 26, 2016 By Adam Bergh Leave a Comment

Today NetApp has dropped some huge news on the storage industry on Day 1 of NetApp Insight 2016 – An entirely refreshed FAS and All-Flash FAS (AFF) Portfolio and ONTAP 9.1!

There is almost too much in here for one post, but I’ll get to the major highlights:

Introducing the FAS2600 – The next gen mid-range hybrid platform.

NetApp’s new FAS2600 is anything but entry level. The successor to the hugely popular FAS2500 series of hybrid-arrays the FAS2600 cranks the entry-level up to 11. Let’s start with what the platform is really all about – Speed. And lots of it. According to NetApp, this beast will be 3x the performance of the previous FAS2500 line. With these new specs I don’t doubt it:

12 CPU cores: 3x more physical cores compared to FAS2500
64GB memory: ~2x more memory compared to FAS2500
8GB NVMEM: 2x more NVMEM compared to FAS2500

Published performance stats have the FAS2600 pushing over 100,000 4k NFS IOPs and 4GBps. How’s that for entry level?

Published specs are as follows vs. the FAS2552 which it’s replacing:

Let’s take a look at the back of the new FAS2600:

Some of the enhancements here are: Mini-SAS running at 12GB. Two dedicated 10GB SFP+ ports for the cluster network, and four 16GB/10Gbe UTA2 ports.

Take a closer look:

Oh yeah, one more thing. Every FAS2600 comes with 1TB of NVMe on board as FlashCache! Now every system in the field will get all the flashy goodness with some NVMe. More to come on the FAS2600 when I get my hands on one.

Introducing the FAS8200 – The Evolution of Enterprise Hybrid FAS

Gone are the days of having four models of FAS8000 to choose from. NetApp is simplifying their product lineup with the single model FAS8200. A three rack unit high dual controller chassis that will scream with performance. Check out the new specs vs the FAS8040, which we expect the FAS8200 to slot into in terms of list price:

~50% performance improvement
32 CPU cores: 2 times more cores
256GB memory: 4 times more memory
Onboard NVMe M.2 flash (Love this)
4 mini-SAS3 ports for external storage
2 dedicated 10GbE cluster interconnect ports
4 UTA2 16Gb FC or 10GbE ports; also supports GbE
2 10GbE Base-T ports
PCIe Gen3 architecture

The published specs vs. a FAS8040

Let’s take a look at the back of the FAS8200

Some of the enhancements here are the four mini-SAS connections, 2 dedicated 10GBe ports for the cluster network and the 4 16GB FC/10Gbe UTA2 ports.

My thoughts:

This is going to be the most popular controller for the mid-range and most enterprise customers. They shaved 3U off the 8040/8060/8080 which is always nice. The performance on this box is going to be very competitive in the market.

Introducing the FAS9000 – AKA “The Beast!”

Ok, now for something totally unexpected from NetApp – An entirely new platform unlike anything we have ever seen! The FAS9000! Whoa, where do we begin on this thing. It’s an entirely new concept from NetApp. A completely modular chassis where every component can be swapped out individually. That means the the CPU controllers have been entirely decoupled from I/O modules and NVRAM.

Here are some of the highlights:

Single model: FAS9000
Modular RAS focused architecture
I/O slots separate from controllers
Individually pluggable and serviceable I/O modules
Hot-swap capable after ONTAP® enhancements
Based on latest Intel Broadwell 18-core processors
Uses NVMe-connected SSDs for onboard flash
Up to 50% performance increase over FAS8080 EX
Enhanced I/O capabilities, higher throughput adapters
40GbE, 32Gb FC
Full 12Gb SAS connectivity to external drives
In-band ACP support
~50% performance improvement
72 CPU cores: 80% more cores
1024GB memory: 4x increase
DDR4 memory
64GB NVRAM: 2x more NVRAM
High-speed HA interconnect, 80Gb bandwidth
Onboard NVMe-attached SSD flash
Enhanced I/O layout
I/O slots outside processor module
PCIe Gen3 slots, X16 wide lanes for each I/O slot
10 I/O slots per controller (plus 2 NVMe slots)

Take a look at the back of the FAS9000

Front View

Here are some specs of the FAS9000 compared to the previous top end FAS8080. A nice bump in all categories. This should satisfy the largest of enterprise workloads.

Let’s talk performance

Let’s compare the new gear’s performance vs the generation be replaced:

I think these numbers speak for themselves. NetApp is really pushing the bar upwards on what a hybrid-array is capable of. Nice Work NetApp.

More NetApp Innovation – 40GBe and 32GB FC

On the heals of becoming the first storage vendor to offer 15.3 TB SSDs in their All-Flash Arrays, NetApp is first to market with 40GB Ethernet and 32GB FC Support!

32GB FC Notes:

NetApp is launching 32Gb FC target adapter in ONTAP 9.1
Will be supported in currently shipping FAS/AFF80x0 controllers
Also supported in new FAS & AFF platforms (FAS8200, FAS9000, AFF A300, AFF A700)
Provides FC target connectivity for end-to-end 32Gb FC installs
FC Initiator support a possible candidate for a future ONTAP release
Shipped with two short wave length SFP+ optics included

40GBe Notes:

NetApp is launching 40GbE adapter in ONTAP 9.1
Supported in new FAS & AFF platforms
FAS8200, FAS9000, AFF A300, AFF A700
Provides end to end 40GbE support from host to storage backend
40Gb to 40GbE connection or fan out to 4 x 10GbE connections
QSFP+ Passive Copper and extended Short Reach(eSR) cables

Next Gen All Flash FAS –A300 and A700

Today NetApp has also announced new distinct naming conventions for the “AFF” All-Flash hardware lineup – The A300 and A700

The A300 is essentially the FAS8200 and the A700 is the FAS9000 with new bezels.

A300 Front View

A700 Front View

The hardware specs match the FAS8200 and FAS9000 systems.

Let’s take a look at the performance of this hardware in the all flash configurations:

NetApp is estimating that the A300 can produce 50% higher throughput that an AFF8040, and the A700 100% improvement in IOPs at half the latency of a AFF8080. Not a bad step up in performance.

ONTAP 9.1 – What’s New

Stay tuned for more blog posts on a full breakdown of what’s new in ONTAP 9.1. For now here are some of the key highlights in the 9.1 release:

Update: ONTAP 9.1RC1 now available to download from NetApp Here: http://mysupport.netapp.com/NOW/download/software/ontap/9.1RC1/

Along with support for the new gen of hardware are two huge new features – Volume based encryption and “FlexGroups”.

NetApp Volume Encryption

Finally! Encryption on FAS and AFF arrays that do not require special disk drives!

Key Features:

Software encrypt any volume, any disk, any system
FAS, AFF, and ONTAP® Select
Lower cost: no need for hardware self-encrypting drives
More granular: volume level
Onboard key manager
Leverage storage efficiency features
Deduplication, compression, compaction
Future proof encryption solution
Software updates will keep the algorithms up to date

Some Additional Note on the new Volume Encryption Feature:

Unique protection for every data volume
• XTS-AES-256 data encryption key per volume
• Node root and SVM root volumes are not encrypted

$0 license required to enable NVE
• Encryption keys are stored and protected by Onboard Key Manager (OKM) – OKM included with ONTAP 9 at no cost

Supported platforms:
• AFF8000 and new AFF platforms
• FAS8000, FAS6280/90, and new FAS platforms
• Encrypting an existing volume
• Re-keying and decrypting a volume is done via a volume move operation
• FIPS 140-2 Level 2 compliance
• NetApp Storage Encryption (NSE) systems and external KMIP server still required

NetApp FlexGroups: massively scalable, high-performance data container

If you’re familiar with NetApp’s seldom used feature called “Infinite Volume” you will instantly recognize what NetApp is trying to do with FlexGroups. FlexGroups basically fix all the shortcomings of Infinite Volume and add a whole lot more to the picture.

The goal of FlexGroups is to create a Single Name Space that meets the following goals:

Linear scale for performance and capacity
Scales to 20PB and 400 billion files
Operational simplicity
Single mount point with automated load and space distribution
Consistent high performance
Predictable, consistent low latency
All-flash containers
Leading resiliency
ONTAP® nondisruptive operations
Flexibility to scale as you grow
6x performance gain
50% lower cost compared to nearest competitor
Inline efficiency guarantees with AFF

FlexGroup — Target Industries and Workloads

Software Dev
Software build environments
Product design
Product Dev/PLM
Life Sciences
Semiconductor logic design and physical design workflows
Seismic processing and interpretation
Rendering and post production
Genomics scanning/ingest, sequencing, interpretation/visualization

More to come of this feature as we get our hands on it in our labs.

That’s it for now folks! Stay tuned for the latest updates coming out of NetApp Insight 2016!

SolidFire Set To Change The Way We Buy Storage

June 2, 2016 By Adam Bergh 1 Comment

During today’s analyst event in Boulder, Colorado, the home of SolidFire, there was some big news dropped on the storage industry.

SolidFire has announced a plan they are calling “FlashForward”. The “FlashForward” program is essentially a capacity based licensing model for the SolidFire software, designed to separate out the capital spend on hardware and software. SolidFire has grown up in the service provider space and as such has seen unique trends in the industry. Gone are the days of licensing being tied to hardware. Gone are the days of rigid pricing models that don’t fit the on-demand nature of today’s businesses. The whole idea here is to offer supreme flexibility in a dynamic allocation pricing model.

To quote Dave Wright, SolidFire Founder, VP & GM: “The days of ridiculous storage hardware prices combined with software licenses forever bound to a specific piece of hardware are over. Customers want a fundamentally more flexible and fair way to procure storage in the next generation data center. FlashForward delivers our customers a more agile, cost-effective and future-proofed approach to buying storage relative to the unnecessary burdens imposed by legacy storage purchasing models.”

The key components of the FlashForward program that help contribute to this substantial cost savings include:

● Fully Perpetual & Transferrable Licensing – no need to rebuy software when upgrading or replacing hardware

● Granular Capacity Licensing – greater alignment of software license purchases with actual usage

● Provisioned Capacity-Based Pricing – eliminates the cost uncertainty stemming from unpredictable data reduction rates and the reliance on high data reduction workloads to make All Flash Storage cost effective by selling software licenses based on provisioned capacity

● Pooled Capacity Model – Software licenses are enterprise-wide, eliminating stranded capacity and allowing for seamless repurposing of hardware to new sites as business needs dictate

● Built-In Economies of Scale – built-in capacity-tier discounts ensure pricing goes down as deployed capacity goes up

This is really going to be game changing for the segment of storage buyers that are creating dynamic storage environments and want to break away from the traditional model of storage purchasing.

New Product Announcements

Image result for SF19210

Today SolidFire is announcing an all new storage node as well: the SF19210 Node

Here are some of the juicy specs of this new beast of a storage node: Twice the performance and capacity in the SF Series – and lowest cost per GB/IOPS. Up to 40-80TB effective capacity, 100,000 predictable IOPS and sub-millisecond latency.

This new node is going to be well suited for those huge database application needing massive IOPs and throughput.

SolidFire is also announcing version 9 of Element OS – Code Named Fluorine

● Integration with VVols – By integrating SolidFire’s Quality of Service (QoS) with VMware VVols, customers can achieve the most granular control over storage performance on a per-virtual machine (VM) basis. This functionality allows them to set min, max and burst IOPS levels, ensuring exact amounts of capacity and guaranteed performance for even the most performance-sensitive VMs. Both capacity and performance can be changed on-the-fly without migrating data or impacting performance.

● Increased Fibre Channel (FC) scalability – Element OS 9 doubles the number of FC node connectivity in a SolidFire cluster to four, allowing users to scale out to 40 storage nodes. Fluorine also increases the limit of IOPS per FC node from 300,000 to 500,000 IOPS, resulting in 2,000,000 IOPS per a four-FC node cluster — over three times better aggregate performance.

● Expanded VLAN features – New features within Fluorine allow for tagging default networks and supporting more flexibility in multi-tenant networks by enabling customers to use overlapping IP addresses on VLANs via separate routing tables, preserving limited IP address resources.

● New user interface (UI) – Re-built from the ground up, the new UI consolidates the storage system data into a single dashboard, saving customers time and resources.

SolidFire’s Field CTO Val Bercovici has this to say about the OS: “Element OS 9 (Fluorine) is a demonstration of our commitment to continuously improving and building an innovative, solid and flexible all-flash array for a broad range of use cases. By simply upgrading to Fluorine, customers gain access to a new set of features that help them transition to a service delivery infrastructure,” “Our unique software-based approach to storage allows customers to enable the next-generation data center in every data center and achieve what’s never been possible before.”

NetApp Announces ONTAP 9! All Release Info Here!

May 30, 2016 By Adam Bergh 5 Comments

Update: You can now download ONTAP 9 GA from here: http://mysupport.netapp.com/NOW/download/software/ontap/9.0/

There’s so much here, so let’s start with the obvious first – The Name Change. NetApp’s flagship OS “Clustered DataONTAP” is now simple just “ONTAP” with the new release to simply be known as “ONTAP 9”. What’s also a bit different is that NetApp will not refer to ONTAP as an OS, but rather just “NetApp® ONTAP® software.” This has a lot to do with the upcoming release of ONTAP Select, the version of ONTAP that will run on white box hardware (more on that later). Other slight name changes include the renaming of NetApp Cloud ONTAP to just “ONTAP Cloud.”

Here’s a chart comparing the old and new names:

What’s Not Changing:
Technical references to specific Data ONTAP products—including clustered Data ONTAP, Data ONTAP 8.x, and Data ONTAP operating in 7-Mode.

So why the name change and why now? In a word, simplicity. NetApp is heading in the right direction here. NetApp has never been known as the array that’s the most simple to use and operate, but you’ll see that with ONTAP 9 and all utilities going forward this will be a major focus for NetApp and it starts with the name change. A good portion of what’s new in the ONTAP 9 release is to simplify deployments and operations.

Let’s talk about new features in ONTAP 9

This is not meant to be a full list, just some of what I think is really cool.

1. New RAID format: RAID-TEC (Triple Erasure Encoding). This new RAID format is designed for triple parity protection for the larger SATA and SSD disk sizes. This will essentially allow for much larger RAID Group sizes on the large disks. Translation: More usable space and better protection of large drives sizes. RAID-TEC will be the default going forward on any drives 6TB and larger and required for drives 10TB and larger. You will be able to covert existing RAID-DP aggregates non-disruptively to RAID-TEC groups as well. Keep an eye out for an upcoming blog post for more in depth breakdown of RAID-TEC.

2. New inline data reduction technology: Compaction. From the company that pioneered data dedupe and compression on tier 1 data, comes a brand new technology from NetApp: Compaction. So what the heck is Compaction? First of all, NetApp currently has inline dedupe and inline compression. Compaction works inline in conjunction with dedupe and compression. The idea of compaction is that storage space usage can be considerably reduced if multiple I/Os or files can be stored together in a singular 4KB block. Compaction happens on logical blocks as they are being organized before being written to storage. NetApp claims this process uses very little CPU and has no performance impact. I’ll detail the space savings on existing volumes as soon as I have a chance to try out this new tech.

3. Onboard Key Manager for Full Disk Encryption at rest! Making full disk encryption more easily accessible is always a good thing. One of the most welcome features of ONTAP 9 has to be onboard key management. This feature was designed to quickly and simply deploy encryption at rest technology. Only a single passphrase is required to set this up. All keys are automatically generated and stored locally in the ONTAP cluster. Most importantly, no need to manage encryption keys! Of course, external key management (KMIP) servers are still supported as needed, but I’m digging this new feature.

4. Advanced Data Partition Enhancements – NetApp has put a ton of work into enhancing the “ADP” feature that was release in ONTAP 8.3. If you recall, the idea of this feature was to return a ton of usable space that was locked up in the required root aggregates in ONTAP. This ended up being a huge improvement in usable space from RAW. The new enhancements in ONTAP 9 take this idea and expand on it further, which is now even more critical with the larger SSD drive sizes hitting the street. NetApp estimates that the changes in ADP will return 17% more usable space in 24-drive configurations vs 8.3.x implementations. Not bad at all.

Other Enhancements:

SnapLock returns. If you remember this feature from the 7-mode days, this gives storage operators WORM capabilities and greater compliance with data retention requirements by being able to lock data in place and guarantee that it has been unaltered.

All Flash Performance Enhancements. It seems that every release from NetApp bumps up performance on all-flash configurations, and ONTAP 9 is no different. NetApp claims that ONTAP 9 should produce 60% more IOPS versus 8.3.1 code.

Faster take over and give back performance. Planned and unplanned failovers should be much faster in ONTAP 9 vs. previous releases. NetApp is claiming ONTAP 9 should have these events down to 2-15 seconds.

Enhanced Analytics. Keep an eye out for a new “head-room” metric that works in conjunction with the upcoming OnCommand Performance Manager 3.0. The idea of this new feature will be to show how “utilized” each node in a cluster is. This provides insight into how much additional workload your system is able to handle.

System Manager Enhancements. The evolution of the on cluster System Manager software continues with ONTAP 9:
- New menu placements
- Real-time performance graphs on any object in the cluster are some of the highlights

ONTAP Select – This one is going to require a whole dedicated blog post for, so keep an eye out for that. In a nutshell this is a full ONTAP 9 single node or 4-node cluster that runs as virtual machines on any hardware. This is the full goodness that is ONTAP with the full flexibility to run on any commodity hardware! This is certainly an evolution of ONTAP Edge, but EDGE was never multi-node clusters with HA. More to come on this.

Hardware Announcements

With all that is new with ONTAP 9, let’s not forget about all the new hardware announcements as well.

15.3 TB SSD – Yes, you read that right. Don’t let the start-ups let you think they’ve cornered the market on innovation. NetApp is first to the market with 15.3TB SSD in a 2.5” form factor. Imagine a single 2 rack unit high – 24 drive shelf full of 15.3TB SSD drives. NetApp is now estimating a 4:1 effective capacity on All-Flash FAS running ONTAP 9. If those numbers hold, a single shelf of these drives would yield over 1PB effective capacity!
New 12GB SAS Disk Shelves –With shiny new SSD drives we’re going to need some new SAS shelves, correct? In step the new 12GB SAS drive shelves. Take a look at the new drive shelves being announced below. Historically the last digit in the NetApp shelf ID has been the SAS speed of the shelf. So why “C” you ask? “C” just happens to be “12” in hexadecimal.

Want to try on ONTAP 9 for yourself? Neil Anderson over at FlackBox.com has produced a free ebook on how to set up an ONTAP 9 simulator on your workstation to try out all the lovely ONTAP goodness. You can get the free ebook HERE. Thanks Neil!

NetApp Releases the EF560 All Flash Array

January 27, 2015 By Adam Bergh 1 Comment

Speed freaks rejoice! Your calls for even more massive speed have been answered!

Today NetApp announces an update to the massively popular EF550 All-Flash Array – the EF560, and what’s in store for your IOPs hungry apps is nothing short of incredible!

The EF560 builds on the proven enterprise reliability and performance, and density of the previous gen EF540 and EF550 models, and takes it to a whole new level of performance.

For those out there reading this that don’t know much about NetApp’s E-Series line of storage, here is a quick refresher:

E-Series arrays block only (ISCSI and FC) performance and density optimized arrays.
E-Series arrays run the “SANtricity” operating system, not to be confused with Data ONTAP, NetApp’s unified protocol OS.
They are designed for speed and simplicity of operation. See my post on how to Deploy One in 30 Minutes
They employ several RAID types, but most notably the “Dynamic Disk Pool” feature, which takes all the complexity out of setting up and managing an array’s potentially daunting amount of disks.

On to the juicy details of the EF560 AFA!

As you can see from the below details, we have similar specs to the previous gen EF550 array, with the added addition of the much needed 12GB SAS and 54GB Infiniband connectivity options.

But what’s really the meat of this release is SPEED, and lots of it! Check on these performance test results:

Yes people, this is 650,000 IOPs at sub .8ms latency! What’s notable here as well is that there are little to no difference in the performance of the DDP RAID type vs. RAID-5 and RAID-10!

Check out some Oracle performance numbers:

On OLTP Databases, the EF560 has 2.5x improvement over the previous gen EF550!

Overall, this is a major performance lead forward for this All Flash offering from NetApp. This is a powerhouse monster that is going to turn a lot of heads and leave more than a few competitors scratching their heads.

Update: The Storage Performance Council (SPC) has released the results from the EF560 – Spoiler Alert! They are pretty amazing. You can download the full report here: http://www.storageperformance.org/benchmark_results_files/SPC-1/NetApp/A00152_NetApp_EF560/a00152_NetApp_EF560_SPC-1_executive-summary.pdf

Why Your Future Data Center will be “Converged”

August 18, 2014 By Adam Bergh Leave a Comment

Unless you’ve been living under a rock the past decade you know that the data center server footprint has been shrinking dramatically. With the advent of server virtualization and individual servers that can now have a TB or more of RAM, it is increasingly common to see what was once rows of traditional servers condensed into a single cabinet. CFOs love the reduced capital costs that server virtualization provides; CTOs love the increased asset utilization.

However, the “miracle” that is data center server virtualization is not without increase: rises in storage and network administration usually follow. In other words, a reduced physical footprint does not necessarily equate to a proportional reduction in IT management personnel. Internal company resources are now demanding faster, more complex, and more automated deployments of resources. In fact, data center virtualization has effectively introduced a whole host of new challenges that include not only the typical management of hundreds if not thousands of easily deployed server and/or desktop operating systems, but also the overlaid applications and physical hosts, along with their corresponding network and storage requirements.

The modern datacenter is a software-defined datacenter; in today’s computing era, it is not only common but expected that a company will make full use of one or more virtualization technologies.

What is “Converged Infrastructure?”

IDC defines converged infrastructure as “the development of a data center (or closet) based on a set of standard elements (e.g., x86 processors for compute, PCI for the system bus, and 10GbE for internode and storage connections). By leveraging hypervisor and storage virtualization software, these “chunks” can then be dynamically partitioned and automatically rebalanced to support a pool of different business applications.”

In simple terms, it’s the marriage of storage, networking, and compute into an elegant, flexible, unified solution. Think if it as an infrastructure building block.

If you haven’t heard of who the major players are in the Converged Infrastructure world, check out IDC’s data below. As you can see this is a trend that is accelerating rapidly and is no small potatoes idea.

Top 3 Vendors, Worldwide Integrated Infrastructure, Q1 2014 (Revenues are in Millions)

Vendor	1Q14 Revenue	1Q14 Market Share	1Q13 Revenue	1Q13 Market Share	1Q14/1Q13 Revenue Growth
1. Cisco/NetApp	$268.4	22.7%	$179.2	25.7%	49.8%
2. VCE	$254.3	21.5%	$176.8	25.4%	43.8%
3. EMC	$179.9	15.2%	$89.9	12.9%	100.2%
All Others	$477.4	40.5%	$251.0	36.0%	90.2%
Total	$1,180.1	100%	$696.8	100%	69.4%

Source: IDC Worldwide Integrated Infrastructure & Platforms Tracker, June 26, 2014

As you can see the Cisco/NetApp collaboration called “FlexPod” is the current market leader. For a quick primer on what FlexPod is, check out this video below:

FlexPod Intro via NetApp’s YouTube Channel

What Are the Benefits of “Converged Infrastructure”

By deploying a converged infrastructure such as a FlexPod that contains unified storage, servers, and networks, an IT organization will see the following benefits:

Reduced risk with proven validated designs from the industry leaders in converged infrastructure.
Massive boosts to operational efficiency and agility to meet key business initiatives.
Lower cost per user without sacrificing scalability
Improved consolidation of storage, server and network resources through the use of 10Gbe, FCoE, virtualization, and highly integrated storage.
Cost reductions through the seamless ability to migrate data resources (servers and storage) non-disruptively
Enable faster, more reliable recovery of data and applications.
Reduce risk of data lost and associated costs.

Considerations in Moving to a “Converged Infrastructure”

For many organizations, moving from a traditional data center architecture to a converged model requires the right partner to help you achieve a successful rollout into the new future of your data center. Netech Corp is an industry leading in FlexPod converged solutions and has many years of successfully deploying this solution in all market verticals.

Check out some references from Netech customers who have implemented FlexPod solutions.

Authenticom: