Nimble Storage all-flash at entry level price
Nimble Storage this week unveiled a new entry-level, full-featured, all-flash storage array with a starting street price of under $40,000.
Nimble Storage also said it has completely refreshed its hybrid flash array line to include significantly higher performance and lower cost, said Gavin Cohen, head of product marketing for the San Jose, Calif.-based company.
The release of the company’s new entry-level, all-flash array comes just six months after Nimble Storage, best known for its hybrid flash array solutions, unveiled its first all-flash arrays.
Since that launch, all-flash arrays accounted for 12 percent of the company’s total array revenue, Cohen told CRN. He could not yet comment on growth in this quarter because Nimble Storage, as a public company, has yet to report the figures.
Nimble Storage’s new AF1000 can scale to up to 165 TBs of flash storage capacity in a 4U enclosure, or up to 2 petabytes effective capacity and 350,000 IOPS in 12U of space, Cohen said.
The starting end-user price of under $40,000 includes 11 TBs of raw capacity, three years of support, and the full range of storage services Nimble Storage provides for its other systems.
The new model includes an all-inclusive software license with flat support pricing and no forklift upgrades required. Customers can also purchase Nimble Storage’s Timeless Storage package, which offers customers a three-controller upgrade after three years.
As the cost of flash storage continues to fall, it is possible that all of a customer’s storage will go to all-flash arrays, said Alan McDonald, president and CEO of All Connected, a Simi Valley, Calif.-based solution provider and long-time partner of Nimble Storage.
Nimble Storage Gets Aggressive on Entry Point Pricing with New Predictive Flash Array
New All Flash Array Offers Superior Scalability Starting under $40K USD with 3 Years of Support; New Generation of Adaptive Flash Arrays Deliver Up To 2X Performance Improvement and 40% Lower Cost of Capacity
Nimble Storage (NYSE: NMBL), the leader in predictive flash, today announced an aggressive entry point to all-flash storage. The new AF1000 is a full-featured all-flash array that offers superior scaling. Customers can start small and scale non-disruptively up to 8PB at a substantially lower cost. Nimble also introduced a new generation of Adaptive Flash arrays with significantly improved price-performance and price-capacity benefits across the portfolio. All Nimble arrays include InfoSight Predictive Analytics to predict and prevent storage and non-storage issues across the infrastructure stack.
The starting point for an AF1000 has up to 20TB effective capacity and is expandable up to 165TB effective capacity in 4U. It can also scale-up and scale-out non-disruptively to achieve over 1.2M IOPS and 8PB of all-flash capacity. Because every Nimble All Flash and Adaptive Flash array runs the same NimbleOS, management and functionality is identical and arrays of any type can be clustered and managed as a single entity in a Unified Flash Fabric. Customers who use All Flash for primary applications can leverage Adaptive Flash for backup, disaster recovery, and archival at a third of the cost of All Flash.
Nimble Storage Price-Performance Advantage
Nimble CS-Series Adaptive Flash arrays deliver flexible performance and capacity for mixed mainstream workloads at one-third the total cost of ownership of legacy hybrid arrays. Built with the same NimbleOS, they can be clustered with Nimble All Flash arrays and managed as a single entity in a Unified Flash Fabric. In this way, applications can be transparently migrated between All Flash and Adaptive Flash.
The new portfolio consists of the CS1000, CS3000, CS5000 and CS7000 and delivers up to 2X performance improvement and 40% lower cost of capacity over the previous generation of Adaptive Flash arrays. The new high-end CS7000 array achieves up to 230K IOPS and scales to over 2.4 PB of effective capacity. Nimble All Flash arrays already benefit from inline variable block de-duplication and compression. De-duplication is planned for the Adaptive Flash arrays in a future release at no cost for existing customers who maintain an active support contract.
The new Adaptive Flash arrays are designed to suit a broad range of primary applications. They can also be built in highly cost-optimized configurations to achieve a preferred platform for secondary storage. When primary applications are on a Nimble All Flash array, an Adaptive Flash array can be used for backup, disaster recovery and archival at a third of the cost of All Flash. The new generation of Adaptive Flash arrays extend this capability further by delivering better price-performance for these secondary workloads.
Each Nimble Storage array includes InfoSight Predictive Analytics to predict and prevent issues across the infrastructure stack. The customer support experience is transformed because tasks handled by traditional Level 1 and 2 support staff are automated. The Nimble support organization is made up entirely of Level 3 experts that answer calls in less than a minute on average. The most complex issues are rapidly resolved because InfoSight already has the information required to solve the problem.
Timeless Guarantee Applied to Entire Predictive Flash Platform
Timeless storage is an innovative new business model built on trust and transparency. Every Nimble array comes with all-inclusive software licensing, guaranteed flat support pricing for the life of the product, no forklift upgrades and a seven-year lifespan on SSDs. There is also an option where after three years, a customer receives a free new controller that is guaranteed to be at least 25 percent faster when they renew support for an additional three years.
The AF1000 MSRP starts under $40,000 USD including three years of support.
For years, storage companies have touted “storage as a utility” and “storage on demand”. The efforts were meant to combat the dreaded “forklift upgrade” of legacy iron which was both disruptive and expensive.
Nimble now weighs in with it’s on-demand option with a published estimate of 6.8 cents/GB per month.
Nimble Storage On Demand
Cloud adoption is growing at an exponential rate. But organizations still have few options when it comes to data storage. Buying storage outright is expensive; leasing typically requires a long-term commitment. Neither offers the flexibility most organizations need in today’s dynamic marketplace. Organizations have been waiting for a storage solution with cloud-like agility and flexibility.
Nimble’s Storage on Demand is a subscription service for on-premise storage that adapts to businesses’ changing needs. It allows customers to decide where their storage will reside – whether it’s their own datacenter, a colocation facility, or the cloud. The storage is configured based on service levels, with billing based on usage rather than fixed configuration. Finally, customers pay monthly with only a one-year commitment.
Pay Just For What You Use.
Monthly billings accurately reflect storage usage, even when it shrinks. Monthly charges are calculated on a simple fixed rate price in cents-per-GB.
Service-Level Subscription. Storage on Demand offers four performance service levels:
Platinum (fastest), Gold, Silver, and Bronze. In each level, billings are based on actual
Never Run Out of Capacity or Performance
A 10 percent buffer, provided at no extra cost, ensures that capacity will always be available as your usage grows. Additional IOPS are automatically included to maintain performance needs.
No Excessive Lock-Ins
Storage on Demand requires a first-year commitment. Customers then have the option to cancel or continue paying
for services on a month-by-month basis.
Storage on Demand includes all the features and support organizations
need to deploy and effectively manage storage:
Helpdesk support ensures end-user services remain consistency available with proven 99.9997% uptime.
Advanced capacity saving features further reduce storage costs, and increase capacity utilization by as much as 2.5x.
Built-in Data Protection
Efficient snapshots and data replication protect against accidental and catastrophic data loss using minimal capacity and network overhead.
Reduced data center footprint dramatically cuts data centerfootprint, power and cooling costs.
InfoSight™ Predictive Analytics
InfoSight monitors arrays to predict and prevent problems — and offer remedies — before they cause disruption. Infosight predicts, plans, and tracks storage usage and performance to maintain service levels and provide detailed reporting.
People generally like Nimble Storage…they really do. They see the WebEx presentation and they’re immediately sold on the hybrid flash solution and its architecture.
But let’s be honest: Nimble Storage isn’t the cheapest solution out there. For many years now, I.T. professionals have become used to quantifying their need for purchasing storage through a “Cost per TB” model. Well, this is a very good model to use for SANs utilizing hard disk drives (HDD). But when it comes to SSD or flash-based systems (or even for hybrid systems, for that matter), this changes the paradigm of how to quantify (and justify) the cost of a flash-based or hybrid system altogether.
So, if you’re an I.T. professional who handles or oversees the storage portion of your network, it might prove productive to ask these questions (based on some questions posed by a storage colleague of mine) as you plan out your storage strategy:
Storage Administrator / IT Administrator / Network Administrator
- Where are you seeing storage latency within your network?
- Is this latency impacting the performance of some of your more data-intensive applications (VDI, SQL, Exchange, Oracle, etc.)?
- What is the current IOPS limit of your SAN?
- If your SAN cannot meet your needed IOPS limit needed for these applications, where do you need it to be in order to meet demand?
- How often are you having to troubleshoot these applications because of low IOPS performance?
C-Level Manager / Administrator / Network Manager
- Think Business Case and ROI (of course, you already know this…)
- Speed isn’t always important, but for certain applications it does – where are these areas you feel the need for speed?
- Would revenue increase if you were to increase this speed in your storage network?
- If your environment measures transactions, how would revenue be impacted if you were able to increase the amount of transactions by 10%, 20%, 30%, 40%, etc…?
Director / VP / CIO / CTO
- You have a fixed budget amount for storage each year: What if you could upgrade the performance of your database applications (Oracle, DB2, Exchange, SQL) with just the budget you have?
- Have you considered the benefits and tax advantages of leasing / financing new storage purchases?
The bottom-line is that organizations need to think of flash and hybrid storage systems as solutions that will help increase the performance of their network and applications which, in turn, will increase the profitability of your organization. Move away from that “Cost per TB” model. That doesn’t work for SSD and flash. Nimble pricing might seem a bit expensive at first, but after you build that business case you’ll never go back to the “Cost per TB” model again.
We are also seeing a trend toward “service level pricing” for many of the large storage vendors; pay for what you use. More on that later.
There are several Nimble Storage MSRP Price Lists floating around. The following Pricing chart is outdated but may still help with budgeting.
Nimble Storage Adaptive Flash
The Nimble Storage Adaptive Flash platform provides enterprise IT organizations a single architectural approach to dynamically tune storage resources to the needs of demanding workloads. Adaptive Flash is based on CASL™, Nimble Storage’s patented Cache-Accelerated Sequential Layout architecture, and InfoSight™, the company’s automated cloud-based management and support system. CASL allows performance and capacity to be scaled seamlessly and independently to accommodate efficient data center growth. InfoSight ensures flash will be deployed intelligently to satisfy the fluctuating requirements of enterprise applications, eliminating the wasteful overprovisioning of storage resources.
Adaptive Flash can:
- Provide multiple service levels to eliminate storage silos with a single architecture for all applications
- Grow seamlessly by scaling compute, cache, or capacity independently and non-disruptively
- Minimize risk with integrated SmartSnap snapshots and consistent backups, SmartSecure encryption, and efficient SmartReplicate disaster recovery
- Improve IT productivity with push-button deployment and the proactive, analytics-based support of InfoSight
After years of selling hybrid storage arrays, Nimble Storage has entered the All-Flash sector going head to head with Tegile, Pure Storage and other.
Nimble in the past has sucessfully sold hybrid arrays that mix solid-state drives (SSDs) with mostly hard disk drives. There are four Nimble all-flash models that vary in capacity and performance. The flagship AF9000 scales to 553 TB of raw storage in a 4U array, with two 4U expansion shelves. The midrange AF7000 scales to 323 TB of raw capacity, with two expansion shelves. The AF5000 and entry-level AF3000 hold one expansion shelf apiece. The AF5000 scales to 184 TB, and the AF3000 scales to 92 TB of raw capacity.
Nimble claimed a mixed read-write performance ranging from 50,000 IOPS on the AF3000 to 300,000 IOPS on the AF9000, and up to 1.2 million IOPS per four-node AF9000 cluster.
Nimble customers can scale up to four AF arrays and manage them as a unified cluster. Nimble presumes 5:1 data reduction with its proprietary data compression and deduplication that boosts effective capacity to 2 PB per 12U configuration, and up to 8.2 PB per four-node AF9000 cluster.
The all-flash systems support Fibre Channel and iSCSI block storage. Nimble dubbed this all-flash system the Predictive Flash Platform, which uses the same cloud-based InfoSight Predictive Analytics as the Nimble Adaptive Flash hybrid arrays.
Nimble all-flash arrays use Samsung Electronics’ PM863 Series SSDs, ranging from 240 GB to 3.84 TB capacities. PM863 uses Samsung Vertical NAND flash technology. Nimble operating system (OS) software provides triple-parity RAID to withstand three simultaneous drive failures per AF array without data loss.
Nimble Storage all flash costs and pricing are just starting to become available.
In our last post, “Flash for Cache, just not for Writes” we talked about Nimble Storage’s approach to accelerating read performance using a true SSD Flash cache. Simple, elegant, FAST. But what about increasing write performance?
Write BIG, Write FAST
Nimble Storage solves the disk I/O write bottleneck by once again identifying the problem and applying new, inexpensive technologies that we unavailable 10-15 years ago. The solution is based on what they call Cache Accelerated Sequential Layout, or CASL.
At the core of CASL write performance story is a design structured around Write Optimized Data Layout. When pure throughput is required for spinning media like SATA drives, the best way to push performance is to write in long, continuous streams. Small, random writes spawn head seeks and drastically reduce throughput.
To solve this problem, Nimble Storage’s CASL “collects or coalesces random writes to the array, compresses and writes them sequentially to disks. This results in write operations that are as much as 100x faster than traditional disk-based storage.” The key to CASL is that the system uses DRAM to facilitate the collection and consolidation of data which is extremely quick.
All this adds up to a relatively inexpensive array that competes head-to-head with gear twice the cost.
I wish there were more complexity to it, but sometimes the smartest ideas are simple!
As we worked through our discovery process of Nimble Storage and it’s amazing claim that it solves the performance vs. capacity trade-off, we stumbled upon an interesting concept….not using the Flash reservoir for write caching. Using the flash only for true, read caching instead of an additional, higher performing storage tier seemed foreign until we read the blog post by co-founder Umesh Maheshwari.
Flash memory shines on reads: it reads 100 times faster than a disk. But its performance advantage is much weaker on writes, and its write endurance is much lower than disk’s. Therefore, Nimble OS uses flash only for accelerating reads, aka “read caching”. It uses NVRAM (a DRAM-based device) for accelerating writes, aka “write caching”.
As with all technology trade-offs, emerging companies without the legacy customer base using older techology have an advantage. New concepts and methods are available to newer companies like Nimble that can architect efficient use of advanced multi-processors, DRAM and flash. In Nimble’s case, they have built an efficient DRAM to disk layer that eliminates the need for a middle flash tier. Flash can be used for what it does best; read speed.
See Umesh’s blog post here Write Caching in Flash; a dubious distinction.
Nimble Storage solutions are built on the patented Cache Accelerated Sequential Layout (CASL™) architecture. CASL differs from the traditional bolt-on approach of using flash as a tier. Instead, CASL is designed from the ground up to leverage the lightning-fast random read performance of flash and the cost-effective capacity of hard disk drives. What’s more, CASL incorporates innovative efficiency features, such as in-line variable-block compression, cloning, and integrated snapshots to store and serve more data in less space.
Dynamic flash-based read caching
CASL caches “hot” active data into a flash-based SSD in real time—without need to set any complex policies. This way it can instantly respond to read requests—as much as 10X faster than traditional bolt-on or tiered approach to flash.
Write optimized data layout
CASL collects or coalesces random writes to the array, compresses and writes them sequentially to disks. This results in write operations that are as much as 100x faster than traditional disk-based storage.
CASL compresses data as it is written to the array with no performance impact. It takes advantage of efficient variable block compression and multicore processors. A recent measurement of our install base shows average compression rates from 30% to 75% for a variety of workloads.
CASL allows for the non-disruptive and independent scaling of performance and capacity. This is accomplished by either upgrading the storage controller (compute) for higher throughput, moving to larger flash SSD (cache) to accommodate more active data, by adding storage shelves to boost capacity, or by combining multiple arrays in to a single scale-out cluster—scaling both performance and capacity beyond a single array.
Snapshots and integrated data protection
CASL can take thousands of point-in-time instant snapshots of volumes by creating a copy of the volumes’ indices. Any updates to existing data or new data written to a volume are redirected to free space (optimized by CASL’s unique data layout). This means there is no performance impact of snapshots and snapshots take little incremental space as only changes are maintained. This also simplifies restoring snapshots as no data needs to be copied.
Nimble Storage efficiently replicates data to another array by transferring compressed, block-level changes only. These remote copies can be made active if the primary array becomes unavailable. This makes deploying disaster data recovery easy and affordable – especially over a WAN to a remote array where bandwidth is limited.
Zero Copy Clones
Nimble Storage arrays can create snapshot-based read/writable clones of existing volumes instantly. These clones benefit from the fast read and write performance, making them ideal for demanding applications such as VDI or test/development.