Manageability is another hidden cost to upgrading an existing NAS infrastructure. With falling prices and improved durability making new storage media such as Flash SSDs widely available to boost application performance, many companies are tempted to install Flash at tier zero and expect that it will solve their performance problems, albeit it at a relatively high cost. But installing fast-access storage media solves only part of the problem. IT then has to figure out which applications are best served by the new tier of storage, often having to become an expert in the latest storage media read and write rates and application QoS schemes in order to optimize the utilization of the more costly storage media. In comparison, an Avere FXT cluster has the intelligence to dynamically allocate data to the appropriate storage tier and media based upon both data and access characteristics, which balances the cost/performance equation with no administrative overhead.

At this point controller upgrades are the typical course of action. Controller upgrades are so common that people have become numb to the pain. Let me remind people of the problems that come with this and propose a better way.

Controller upgrades with a traditional NAS system involve the following steps:
1. Purchase higher performance controllers (typically a model or two up the traditional NAS vendor’s product line, if a higher model exists)
2. Purchase new disk shelves or Flash-based PCI cards, required to get more performance out of the new controllers
3. Purchase new licenses for all software (e.g. NFS, CIFS, mirroring) at the higher price tier of the new controllers
4. Take the original NAS system offline
5. Remove the old controllers
6. Install the new controllers
7. Add the new disk shelves
8. Bring the upgraded system back online

The above process is expensive, disruptive, and requires more disks, power, and space even if no additional storage capacity is required. As an alternative, consider boosting the performance of your existing NAS with Avere FXT appliances.

Avere FXT appliances boost the performance of all NAS applications by accelerating read, write, and metadata operations without the need to add more disks to your NAS system. FXT appliances scale from 1 to 25 nodes per cluster to match your initial needs and gracefully scale as your needs grow. If more capacity is needed, then this can be accomplished by adding cost-effective SATA drives to your existing NAS system. FXT appliances are simple to install in existing environments and require no changes to existing applications, clients, NAS systems, or data retention procedure such as backup and mirroring. FXT appliances accelerate NAS performance and enable a 5:1 reduction in disks, power, and space when compared to traditional NAS.

With Avere you get the benefits of a clustered storage solution today without the need to migrate your data to a new storage system. For more information on FXT appliances, please visit the Avere website product page.

Typically, Flash is added to a NAS system in one of three ways, as a PCI card, an SSD array, or a caching appliance. Let’s look at the leading vendor in each category and the efficiency of their approach.

NetApp offers Flash as a PCI card they call PAM. Up to five PAM cards with a total of 2.5TB of Flash can be installed in some NetApp controllers and pricing is in the $170-200/GB range. PAM is inefficient for two reasons. First, PAM is read-only so customers still need many hard drives to handle the write workload. Second, PAM can only help the read performance of the controller in which it is installed. This means you may be adding up to five PAM cards to every NetApp controller in your environment. This approach is expensive and highly inefficient in environments with multiple NetApp systems since Flash is added to individual storage “silos” to handle the peak load on the silo but is under utilized much of the time.

EMC offers Flash as an SSD array. Up 16 SSDs and 6.4TB of Flash are supported per array and pricing is in the $200-300/GB range, making a fully populated array very expensive. Beyond the price, EMC’s SSD arrays are inefficient since data movement to Flash is slow and not granular. Data movement is triggered by a policy engine that measures data activity across long periods (e.g. days) and cannot immediately respond to a hot application. Also, data is moved in large volume-level chunks meaning lots of cold data is using expensive Flash storage alongside the hot data.

Avere offers Flash as a caching appliance that sits in front of NAS systems from other vendors. The Avere FXT 2700 has 512GB of Flash per appliance, scales to 25 appliances (13TB) per cluster, and accelerates read and write workloads. The Avere architecture is highly efficient since it provides a consolidated Flash layer that is shared by all the NAS systems. This allows customers to provision the right amount of Flash to deliver the aggregate performance across all the NAS systems. In addition, the FXT 2700 makes the most efficient use of the Flash storage by moving data in real-time and at the finest level of granularity possible, blocks with files.

For more information about the FXT 2700 and other FXT models, please visit the Avere website product page.

The value proposition becomes even worse when companies over provision capacity and short-stroke drives in a futile attempt to achieve higher performance. Instead of trying to extend HDD technology, NAS users will get a better return on their storage investment by moving from a homogeneous solution to a tiered NAS solution. Adding intelligent storage tiering, that makes the best use of newer solid state disks, produces a dramatic gain in IOPS performance, with far lower impact on power, cooling, rack space and data center floor space. This approach eliminates the expense of adding potentially hundreds of enterprise HDDs to the NAS environment, recovers the capacity on existing HDDs that was previously unavailable due to short-stroking, and solves NAS performance issues while enabling IT to leverage low-speed (and low cost) SATA drives for capacity requirements.

The two-stage tiered NAS architecture, using Avere’s FXT family of products, provides for all of these benefits by permitting the addition of solid state media to any existing NAS infrastructure.

Unfortunately, one of the giants in the industry has even further muddied the waters. According to a TechTarget article last week regarding NetApp’s earnings call, NetApp’s CEO said that “the whole concept of tiered storage is going to go away.” Presumably, this refers to EMC’s Fast technology, since one industry Goliath always needs to beat up on another Goliath. The unfortunate thing for NetApp, is that on the same call, they completely reversed their position on tiering when they touted the success of their form of tiering, which includes their SSD-based performance accelerator. Tiering is dead, long live tiering!

The simple truth is that no single technology has ever proven to be the panacea of data storage. SATA drives have the lowest cost per bit and are great for archival storage. FC or SAS drives offer a compromise of performance and capacity, excelling at large block accesses like those found in large sequentially read files. Solid State Devices, based on Flash, offer unsurpassed performance for random reads and small block sizes.

More importantly, the differences from one technology to another are measured by orders of magnitude, not by mere percentage improvements. Because of this, a solution that can leverage the strengths of all the technologies is guaranteed to out-perform and cost less than a solution that only uses one or two.

Rather than predicting which storage tiers will win and the capacity of those tiers in a solution, the important information needed to judge a tiering solution is how the tiers are used. Most vendors are completely silent on this. Here are three examples of the more egregious mis-steps in dynamic tiering.

The first mis-step is tiering at too large a level of granularity. Consider a solution that tiers at the volume level. If a few files in a volume become active, the entire volume will need to be promoted to a more expensive tier to get the performance needed for the few files. This results in cost inefficiency as extra data is promoted to the expensive storage and performance inefficiency as the entire volume consumes read/write bandwidth of both tiers that are involved in the promotion of the volume.

The second mis-step is not tiering frequently enough. Several vendors have proposed tiering schedules that are measured in terms of days. This is crazy. Consider the file that I am editing for this blog. I might work on this file for a few days and then rarely, if ever, look at the source file again once the blog is posted. If activity is measure across days, by the time this source file is promoted, it should be archived.

The third mis-step is not using the correct media. Most vendors actually completely avoid this question and require the administrator to set policies. In those instances where the vendor does decide, frequently wrong media is chosen.

An example of this third mis-step is to examine the two stage architecture promoted on NTAP’s earnings call – SSD & SATA. In their architecture the SSD-based performance accelerator is apparently only used for read data. All write data is sent to SATA storage. This is terribly inefficient and is even proven in their SPECsfs®08 posting. To achieve the same performance in a NTAP 3160 with the accelerator module, they required almost twice the number of SATA disks than when they run the same benchmark using FC disks (96 SATA disks versus 56 FC disks). Since the SATA disks have over 3x the capacity of the FC disks, they deployed over 6x as much capacity to store the same amount of data. This over-provisioning is a result of not tiering the media properly and is extremely inefficient in terms of space, power and equipment costs.

What is clear from all of the press on “Dynamic Tiering”, is that the term is both extremely overused and misunderstood. Because of the orders of magnitude differences in storage media costs and performance, data storage solutions can clearly benefit from tiering if executed properly.