Storing data in the cloud is on almost every company’s radar screen. However what remains a little hazy is the best way for them to store data in the cloud. The use of public cloud storage gateways as a means to store and retrieve data from public storage clouds is rapidly emerging as the preferred option to do so. In this first blog entry of my interview series with BridgeSTOR’s CEO John Matze, we take a look at some of the different gateways solutions available for accessing public storage clouds and how they differ.
Jerome: John, thanks for joining me today and taking time out of your schedule to speak with me. To kick off our conversation, could you tell me a bit about what BridgeSTOR has been up to and how it is implementing its public storage cloud gateway?
John: Jerome, thanks for meeting with me. To answer your question, BridgeSTOR is building a cloud storage file system (CSFS.) Basically it is a file system that talks REST out the back end and runs natively in RedHat and CentOS. You can then use standard NFS and Samba on top of it to export it out. You now have a cloud gateway that is more open than many of the current cloud gateways that are out there because it runs on top of standard Linux operating systems.
BridgeSTOR has working on CSFS for about two years. The back end has full encryption to the cloud gateways as well as deduplication and compression to optimize data transmissions over the wire plus it also deduplicates and compresses data at rest.
One of the interesting things that BridgeSTOR has seen with a lot of other products in this space (Panzura, TwinStrata, etc.) is they all have a global file system (which BridgeSTOR also has.) The features and benefit that BridgeSTOR offers are going to be equivalent to a Panzura though what BridgeSTOR has done is it is throwing the appliance model out the window and going to more of a SAS (software as a service) model.
BridgeSTOR ships its solution as a virtual appliance which an organization may go ahead and deploy for a small amount down and then they only need to pay so much a month for access to the data. Then an organization can choose whichever go to an Amazon S3, an Amplidata or some other private cloud. It just becomes a cost factor like anything else. But basically it’s a low entry point to get in and a monthly basis.
What we found for the public cloud guys like Amazon is that a lot of the people are staying away from them, primarily because they do not understand the pricing and how they are going to get dinged on a monthly basis. So BridgeSTOR is looking at putting together a flat pricing model where customers can just access their data through a virtual machine which appears as an F drive that shows up, and whatever you copy to this drive is immediately placed in Amazon’s cloud.
BridgeSTOR will take advantage of some disk cache to make sure that the most read files will stay local. But you do not need all of the files locally because some of the data, once it is there, you are never going to touch it again.
Using the global file system organizations will still be able to view all the files as if they are local. That is one product BridgeSTOR has been working on. You’ll see an introduction coming out about that. I think it is going to be very disruptive from the standpoint of the pricing model. A lot of these object stores and public storage cloud products that are out there are not taking off because the gateways are too expensive. People quite honestly do not want to write their own REST interface for accessing these boxes.
BridgeSTOR wanted to make the cloud affordable so people would start using it. This is something I have been wanting to do for a long time, just even for my own company. I’d love to get rid of the storage requirements. At 9.5 cents per gigabyte at Amazon using storage that is 99.99% available, why not? I don’t think Amazon is going to be like Nirvanix and will go away any time soon.
Jerome: My concern about a company like Nirvanix shutting down so abruptly is that do you think is will have a long term ripple effect on people adopting cloud storage? It’s certainly going to raise a lot of the fears that people have expressed over the years or reignite them.
John: You’re right. But everybody I’ve talked to about my particular strategy is do not care who he is on the back end is. BridgeSTOR is going to be a bit like the country of Switzerland. They do not want to use just one vendor on the back end; they want to use two for this exact reason. If one unexpectedly goes away, they have another copy of their data with another public cloud storage provider.
The challenge that BridgeSTOR had when building this product was, “How do you allow people to have basically a drive F be Nirvanix, and a drive G be Amazon, and they could just copy files from one to the other, or copy whatever they want up to the other?“
BrigdeSTOR has been able to pull that off. That is a feature that the current gateway guys cannot do, or cannot do very efficiently. If the gateway is easy with sufficient redundancy, people will start using these solutions.
In Part II of this interview series, John reveals some of the unique interfaces in which BridgeSTOR will offer on its public cloud storage gateway appliance.
In Part III of this interview series, we discuss the business value that a public storage gateway appliance with a VTL offers and how it affords organizations the opportunity to move their tape museums into the cloud.