Before there was network attached storage or “NAS”, there was simply the file server. Has much changed with cloud NAS? Well of course, but…
Cloud storage is an overarching term describing all sorts of cloud storage services: file, block, object, database, etc. Cloud storage ranges from personal cloud storage where anyone with an internet connection can use data storage in the cloud to backup and actively use files.
Yet, if you work for an enterprise, or small business moving massive amounts of data, your storage solution could be scale-out NAS, or “cloud NAS” or a cloud file system. Since Resilio is in the business of delivering files over distance (as well as file access and file sharing), this blog looks at some of the historic trends of network attached storage leading up to cloud NAS, in the context of data movement and remote file access.
Cloud NAS is somewhat amorphous. Your NAS could be a virtual Linux XFS instance you manage yourself on AWS. It could be a virtualized Windows file server (like AWS FSx, Microsoft Azure Files, or Azure NetApp Files) . Depending on your workflow, your cloud NAS could be persistent or temporary (provisioned on demand). In hybrid scenarios, you may use a cloud storage gateway deployed on-premise that moves data to the cloud behind the scenes, out-of-band from user file access.
In some cases (and we hope you don’t) you go cloud Nasty: I.e., live with file servers that are hard to manage, perform poorly, and, depending on the pricing and egress fees, quite expensive to store and retrieve data.
The best cloud NAS or file-based cloud services solution for your company will likely depend on factors such as applications, workflows, and use cases such as disaster recovery. For instance, if you have 2 active data centers running Windows file servers, you may consider a 3rd site running Microsoft Azure Files or Azure NetApp files in the cloud, as a disaster recovery target or remote workflow.
Network Attached Storage Trends
Cloud NAS has roots in traditional network attached storage. Before there was WiFi and high bandwidth broadband connectivity, file servers and NAS devices were everywhere, connected to the LAN. File access was almost always over low-latency LANs.
The use cases are many. Functionality varies greatly among vendors. Some NAS (phyiscal or virtual) are high end, high performance backed by SSDs and/or erasure coded file systems backed by external hard drives. Others are geared toward simplicity for small businesses.
Yes, for all network attached storage, a common problem remains: coherently moving and managing data across silos of NAS. In some ways the cloud helps; in other ways, the cloud simply adds complexity to data management.
Trend 1: File Server Sprawl.
Let’s look back at a few historic trends leading up to Cloud NAS (well before mobile apps had moved from BlackBerry to Apple iOS).
In the early 90’s, I was working at Microsoft using the NT 4 operating system (sitting next to Microsoft’s Apple Mac support team!) and was blown away with how easy it was to setup and deploy file servers. Just as the NAS storage boom was taking off, Microsoft Windows file servers sprouted up across enterprises everywhere, like mushrooms and lichen moss behind Building 1 on Microsoft’s campus.
Microsoft’s initial strategy to manage file server sprawl was the Distributed File System (DFS). To this day, DFS works pretty well. The only real problem was file replication and synchronization. At the time, Microsoft IT managed (likely) the largest worldwide deployment of file servers using DFS. Under the namespace, IT used various ad hoc “follow the sun” file replication strategies using XCOPY and eventually Robocopy. Today, DFS-R can be used to replicate files. But it’s not very useful over WANs and fraught with reliability issues. (FYI: Resilio has an elegant replacement for DFS-R.)
By the late 90’s, some men from EMC wearing dark suits and sun glasses walked on to the Microsoft campus carrying floppy discs containing a notorious “BIN file” to service the EMC Symmetric SAN stored on campus. At the time, the only people allowed to work on EMC products worked at EMC.
By contrast, Dave Hitz, a smart and friendly guy running a burgeoning startup called Network Appliance, or NetApp, knew he could do better than both EMC and Microsoft to win the file server business. Back then, Hitz and his customers called NetApp’s new file servers “toasters”. They were as easy to setup and use as toasters–and high performance.
Cloud File Storage Gateway for Hybrid Clouds
Try Resilio ConnectTrend 2: NAS eats File Servers.
An Ethernet connection to the LAN and file migration tool was all that was needed to migrate Windows file servers to NAS from EMC, Sun, NetApp, et al. Under the banner of file server consolidation, EMC and NetApp feasted on Microsoft’s file server business. NFS came native on EMC and NetApp. And Samba (an open source ported CIFS/SMB 1.0 file serving stack) was used to serve Windows clients. Andrew Tridgell (of rsync fame) started Samba and gave EMC and NetApp a jumpstart on crafting a multi-protocol file sharing stack. To say the least, Microsoft was not happy. I know first hand, since I was there, working on licensing terms for “CIFS 1.0” for SNIA.
Trend 3: Traditional file access and file sharing protocols like NFS and CIFS/SMB were designed to be accessed over low-latency ethernet LANs.
Accessing NFS or SMB over the WAN (wide area network) or via a VPN will likely not perform as well as a LAN, if there is much latency. We do hear some customers are doing it with SD-appliances, with ok results, depending on the use case. However, cloud native workflows where file access is coming from within the cloud, are best suited for file access to over SMB/NFS.
Another option for is to offload file access to a cloud storage gateway on-premise. Resilio Connect, for example, can be installed on a Windows server on-premise and move files directly into the cloud, predictably at high speed, over any distance and network. The Windows Server simply keeps working as an on-premise file server. That way users can maintain their SMB file access workflows. Pretty cool considering options like Selective Sync in Resilio make it efficient moving data, where only needed files are downloaded or sync’d back to the cloud.
Trend 4: Scale-out NAS eats NAS
Sound familiar? By the mid-2000’s, scale-out NAS begins to gobble up traditional NAS. Simply increase storage capacity and performance on the NAS. Thus creating larger, higher capacity, storage silos with more data to manage. Like raking leaves into bigger and bigger piles. You can put more data on scale-out architectures but they still need to be managed as individual piles of data.
The question now becomes: How do we manage multiple, distributed petascale data sets? Do we simply replicate silos of scale-out NAS file systems across locations, and across N-number of cloud storage endpoints? Also, what about situations where remote workers at home have data stored on direct-attached storage or SoHo NAS from Synology diskstation, QNAP, or another device?
The answer for your company could be to use Resilio Connect in combination with your storage vendor of choice.
Trend 5: Cloud NAS!
Well, Cloud NAS may not help you move and manage all of your massive data sets–but Cloud NAS is easy to deploy and access and could complement your various data storage and movement strategies.
That said, the number of cloud NAS offerings on the market today is hard to keep track of. There are on-premise data center storage gateways (like Resilio Connect Gateway for Azure Files) used for file sharing and a variety of cloud NAS offerings. Some enterprises may choose to create and manage their own cloud NAS via various Infrastructure as a Service (IaaS) offerings–ranging from Microsoft Azure Files to Azure NetApp Files to Amazon AWS FSx. All of these solutions have pro’s and con’s, depending on the use cases, applications, and workflows.
What we will cover in our next blog is the ability to cohesively connect islands of storage–from on-premise to cloud–and move data independently of the storage architecture. Do you need a “cloud file system” or Cloud NAS used in combination Resilio Connect? Stay tuned.
Learn more about Resilio Connect and how it can compliment your Cloud NAS. Or, schedule a demo or start a free trial to see for yourself.
