Moving your data and applications to the cloud isn’t the easiest of tasks, if you want to do it right. There’s a multitude of decisions to make. Some you’ll get wrong, which might make you reconsider your cloud operating model or cloud provider. Which brings the next question: are you locked-in at your cloud provider? Can you move your data between clouds?
One start-up that attempts to make the move to the cloud and moving between clouds easier, is Elastifile. An Israeli company, founded in 2013 with its first version of the product out in Q4-2016, it created the Elastifile Cross-Cloud Data Fabric. Their objective: bring cloud-like efficiency to the on-premises cloud, and facilitate a easy lift-and-shift into the hybrid cloud.
Elastifile Cloud File System
At the core of the Elastifile product is the Elastifile Cloud File System, or ECFS for short. Its designed from the ground up to provide a scalable data platform for a new generation of web scale clients and users. That’s 100.000’s of filesystems, with practically unlimited files and folders.
ECFS is a software only and flash only solution, which runs on either physical systems, on-premises virtual machines or public cloud virtual machines. Storage is presented out over NFSv3/v4, SMB2/3, S3 and HDFS.
You can deploy the ECFS nodes in either a hyperconverged (HCI) mode or a dedicated storage mode (DSM), depending on your architecture preferences and existing infrastructure.
Over 60% of the Elastifile customers are service providers, and in the future these service providers will announce their own preconfigured software+hardware solutions. This makes it easier for other Elastifile customers to find a fitting and proven combination of components for their own datacenter/cloud. Elastifile is also thinking of going for a global partnership with one or several hardware vendors.
Licensing is subscription based, with a 50TB initial license based on raw capacity. After that, you pay as you grow, checked on a quarterly basis. Getting your data on the Elastifile Cloud File System is pretty easy as well, using traditional migration tools such as robocoy or rsync.
We’ve had a lot of tiering vs caching discussions the last Storage Field Days. To make things simple, ECFS is only available in an all-flash configuration, so no tiering across flash and HDDs. It also doesn’t cache data, only metadata. Elastifile claims that it’s a waste of memory to cache data on an all-flash system. With media latencies dropping and Intel’s work on bringing software and controller latencies down as well, this should keep latencies in check at a consistent 1-2ms.
My thoughts on Elastifile Cloud File System
During the original Storage Field Day 12 presentations (which you can find here, check them out), I was initially impressed by the ECFS system resiliency. During the demo, failures were handled with minimum impact to performance and admin interactions. I echo that approach, especially in the time where we need to manage bigger (file)systems with fewer admins. Component failures are a fact of life, so we should have to spend as little time as possible in resolving them.
In a follow-up meeting with Elastifile’s Andy Fenselau and Jerome McFarland, the “moving to the cloud should be easy” message was reinforced. Even if you don’t have any plans to move to the cloud tomorrow, ECFS sounds like a solid product for your own on-premises datacenters. This creates the opportunity to later extend the Elastifile Cloud File System into any public cloud. End goal should not just be hybrid cloud, but also hybrid IT: systems spread across multiple sites, with apps and data dynamically moving to the right location. This enables your IT to automatically adjust to changing demands, without too much administrative overhead.
For more information on Elastifile, check out Ray’s post for more insight in the metadata handling, or Chin-Fah’s post.
Disclaimer: I wouldn’t have been able to attend Storage Field Day 12 without GestaltIT picking up the tab for the flights, hotel and various other expenses like food. I was however not compensated for my time and there is no requirement to blog or tweet about any of the presentations. Everything I post is of my own accord and because I like what I see and hear.