November 19, 2014
As you may or may not know, storage has historically been a pain point for most organizations in terms of exploding data growth, data protection needs (disaster recovery), increasingly demanding virtualized applications (want access to read/write very quickly), and the fact that storage technologies have been complex/difficult to manage.
Furthermore, the majority of traditional storage players are not well positioned to address these difficulties because:
- They still use a fixed block methodology
- Their technology requires a flash tier for random writes which results in having to park data in flash thereby chewing up flash space and causing unnecessary silicon wear on the flash. (By inserting flash into a tier, the flash inserted there has to be expensive, high endurance flash that requires RAID protection due to the inherent risk with flash because the data has to park there. Tiering in general also requires the concept of migration which requires excess flash)
- Traditional disk used in their solutions are slow at processing random I/O operations at a rate that today’s virtual applications require
- They have had to address the need for flash in solutions by bolting on a tier of flash into their solutions which provides a weak foundation architecturally, an outdated file system that can’t deliver the real performance promised and it’s just proven to be unreliable/expensive
- Overall their technologies are based on legacy storage tiering concepts that are really 15 to 20 years old
Nimble Storage looked at all these pain points and saw an opportunity for a ground-up redesign of storage technologies that leveraged the best of both flash and disk. Some of key decisions that Nimble asked themselves during the design process were:
- How do we leverage flash for performance with persisting the data on flash given the endurance concerns?
- How do we leverage disk for capacity?
- How do we be media agnostic and flexible as the relative merits of flash and disk evolve?
- How do we leapfrog incumbents on data management by achieving efficient management without all the costs?
After looking carefully at each of these questions, Nimble created the CASL File System which stands for Cache Accelerated Sequential Layout.
Why is Nimble Storage better from a data protection perspective?
Nimble uses a high efficiency snapshot methodology that is pointer based which has the following inherent benefits:
- Removes backup tier from equation as the data tier as is also the backup unless traditional storage players
- Ability to replicate snapshots extremely efficiently
- Ability to retain snapshot data for long periods
- Ability to replicate data over a modest amount of bandwidth
Fact – A large majority of traditional storage player customers don’t even use snapshots due to the fact that they are copy based and chew-up way too many resources and capacity.
Why is Nimble Storage better from a data replication perspective?
- All Inclusive Licensing Model – Customers with (2) Arrays automatically have replication
- Snapshot based replication method only sends changed blocks = extremely bandwidth friendly
- Result in being able to replicate over WAN link
- Due to efficiency there is no penalty to capacity for performing or keeping a high volume of snapshots
- Granular – Replicate on a per volume or volume collection basis
- Extremely Flexible Replication Capabilities – Select a specific business critical APP with 15 Snapshot Intervals with WAN Replication = Good Nights Sleep!
- Couple with VMWare SRM = Complete and Total DR Solution
- Data Replication Options:
- Primary to DR – 1 WAY
- Many to 1 – Remote Offices replicating to a Central System. Central system does not need to be very big to sustain retention from multiple remotes
- Bidirectional – Becoming more prevalent. Multiple sites that are production so that each site is the other’s DR or the others business continuity failover site
Why is Nimble Storage better from a scaling perspective (performance or capacity)?
- Capacity and Performance are Decoupled from Each Other. Adding Capacity doesn’t modify performance and vice versa
- Capacity – Scale Deep
- Non disruptively add Storage Shelves to Array
- Attach (3) Shelves to a Single Nimble Controller
- Only takes up 12RU and provides 250TB of Capacity!
- Performance –
Alter the amount of CPU by swapping out controllers- Modify the amount Flash CACHE/SSD
- Results in ability to sustain a large active data pool
- Ability to tie multiple independent arrays together utilizing existing Architecture
- Allows customer to manage all arrays together via single user interface
- No Special cables to tie arrays together – iSCSI
- Can have 4 nodes in a group and mix of Nimble models
- Can pool performance resources to increase performance capabilities
Traditional Players:
- Limited scale up or scale out possibilities. Typically can only scale in terms of spindles where as CPU is typically where it’s most needed
- Results in expensive and difficult scaling processes
Nimble:
- Scaling was a key foundation of original design. Address either performance (flash,cpu) or capacity (disk)
- Results in all scaling being non-disruptive and non-production impacting
Why Nimble Storage uses variable blocks vs. fixed blocks?
Traditional Players:
- Use fixed blocks with no inline compression
Nimble:
- Uses inline compression with variable blocks. Variable blocks facilitate compression resulting in 30 to 70% less disk and flash resources used
How Nimble Storage uses flash resources vs. other storage players?
Traditional Players:
- Improper use of flash resources due to the migration events that need to occur in moving data back and forth between tiers – results in wasting flash resources and silicon wear
Nimble:
- Uses flash as a dynamic caching layer so the data is always transient and is populated very quickly – results in the ability to use substantially less low cost flash to accelerate reads
How Nimble Storage is differentiating themselves from the competition
Nimble Storage’s disruptive approach to design has created a CASL File System optimizing hybrid storage solutions that successfully reduce many organizations pain points by taking advantage of the best of both disk and flash.
Like what you read?
Contact us today to discuss how you can leverage Nimble Storage in your organization.
About Mindsight
Mindsight, a Chicago IT consultancy and services provider, is an extension of your team. Our culture is built on transparency and trust, and our team is made up of extraordinary people – the kinds of people you would hire. We’ve always prided ourselves in delivering the full spectrum of IT services and solutions, from design and implementation to support and management. Our highly-certified engineers and process-oriented excellence have certainly been key to our success. But what really sets us apart is our straightforward and honest approach to every conversation, whether it is for a local business or global enterprise. Our customers rely on our thought leadership, responsiveness, and dedication to solving their toughest technology challenges.