Accelerate Your Applications on FlexPod with MAX Data
Cisco and NetApp’s FlexPod converged infrastructure continues to provide an ideal platform for delivering groundbreaking technologies and making those innovations safe and consumable for enterprise organizations. I’d like to introduce you to one innovation called ‘MAX Data’ that is proving to accelerate the performance of enterprise applications. I’d like to illustrate the concept with an analogy that hopefully most can relate to – the speed of getting served in a store to the speed of data in an application.
First, some coming attractions…
This fall, a new enhancement from our FlexPod partner, NetApp is coming and already showing increased application performance in the lab. Its an exciting, optional feature called “MAX Data”. MAX Data is an innovative NetApp server-side storage technology that works with Intel Optane. When using Max Data, it presents Optane media as a traditional POSIX compliant file system enabling existing Unix and Linux based apps to work natively with Optane without rewrite. It’s not a cache but it will deliver a persistent storage layer and eventually include integrated data protection and replication to protect application data against faults. So, FlexPod with MAX Data will bring the performance of persistent memory to many of the applications you run today, delivering orders-of-magnitude faster transactions. Check out the NetApp blog for even more details.
To illustrate the basic concept of MAX Data, I offer an example most people can probably relate to… automotive parts stores. As a teenager, I had an after-school job as a sales clerk in an automotive parts store. Customers would come in looking for parts to repair their old vehicles. Depending on the availability of the parts (i.e., the “data”) there were four different speeds (i.e., “latency”) they could expect. Let’s say a customer came into the store looking for an oil-filter for his 1982 Oldsmobile Cutlass (yeah, it was many years ago!). He might expect that exchange to go something like this:
- “Very Fast Service” – The oil-filter is on the store-shelf for the customer to find, grab and leave in 10 minutes. This is like data residing in memory – ready and available at a moment’s notice. Like shelf space though, it’s relatively expensive. It takes valuable real-estate and requires special packaging, display and signage. It’s also more at risk of loss by theft or damage. The same is true for DRAM memory – adding more can be costly and increase the chance for data loss.
- “Pretty Fast Service” – If that oil filter was not on the shelf, I’d have to help the customer look it up and go find it ‘in the back’. Pretty fast, but it required an extra step, was a little cheaper for the store than valuable shelf space, and, the parts were protected behind the counter. This is like NVMe flash storage – close by and an extra step in the process, slowing things down a little. But ‘persistent’ since the parts were quickly available and less prone to ‘loss’.
- “Not so Fast Service” – No filter in stock? In that situation I had to send the customer across town to the local depot which took at least an hour of his time. This is akin to an app having to grab the data from a local hard-drive. Cheaper for the store, but a big ‘performance’ hit — customers were annoyed when we didn’t carry parts in stock. Image having to constantly do this? Having to constantly go back to the hard-drive for data?That’s a pretty inefficient process .
- “Pretty Slow Service” – What if the part was not even at the local depot? I’d have to place an order to have the part shipped across country from a wholesaler in three to five days. This is similar to cold storage in the cloud or even tape back-up (for those that remember those days!). It was the lowest performance option (sorry, no oil change this weekend like you planned, sir) but cheapest for the store, with no inventory cost at all.
MAX Data for Increased Density and Performance
What if you could move as many high-volume auto-parts (i.e. “data”) onto the store shelves (i.e. “memory”) as possible without quite the typical expense? This is a little like the operating model many stores eventually went to in recent years – serve-your-self in a big-box warehouse. With MAX Data, NetApp’s latest testing shows double the VM density and three times the storage performance which is like having less clerks with better productivity.
From the auto parts store perspective, this would provide maximum satisfaction to as many customers (“applications”) as possible, benefiting the store (i.e. the FlexPod within your data center) immensely. It would also result in faster through-put and likely more revenue for the store due to no extra steps and wait-time.
This is like what adding persistent memory with MAX Data does. Max Data is designed to deliver both accelerated performance and data services like adding a sophisticated security system so the inventory (again, your “data”) is protected from fire and theft (i.e., rapid recovery), up to the last transaction, in the event of sudden database server failure.
Two Powerful Technologies Working Together
Is all this available on a FlexPod CVD today?… no. But we are continuing to test and run configurations that are giving some terrific results. Check out our Accelerated Data for FlexPod white paper. That white paper is a living document and we’ll keep updating it as we test out new features, new use-cases and new performance characteristics with this exciting new technology on FlexPod.
Cisco and NetApp’s FlexPod converged infrastructure is working to provide an ideal platform for delivering groundbreaking technologies and innovations that are safe and consumable for enterprise organizations. We’ll be at VMWorld this August, so be sure to come check us out at both the NetApp and Cisco booths for more information.
In late October we’ll be at “Insight”, NetApp’s annual showcase. We’d love to see you at either event and show how using these two powerful technologies can help you with your company’s unique needs.