**Full Disclosure, Used to work for NetApp, providing support for their FAS series appliances.
Even with a huge power-eating, air conditioned server-room-requiring giant enterprise grade hunk of storage-- they are going to find the same problems. SAN devices, by their nature, are going to require a specialized pipe (Typically fibrechannel), and to be efficiently reached over the public internet are going to need a dedicated VPN with link aggregated enpoints. The exact same protocols are going to be used (NetApp actually licenses the exact same code that MS uses in their windows server lineup, so unlike this consumer box, does not use SAMBA-- but it is STILL SMB protocol), such as SMB and NFS, just that in addition to that there are several proprietary ones, like SnapMirror and pals, special interfaces for inter-device clustering for graceful failovers, et al.
There is no magic bullet here. The enterprise grade appliances have some interesting features that work at the block level (instead of FS level), making them useful for things like iSCSI LUNs and the like that most end users would never even dream of using, or for aggregating disk arrays into virtual volumes and the like-- but otherwise, the actual nitty gritty of REACHING the storage is going to be over the exact same protocols, and is going to have the exact same use case restrictions and caveats. They tend to have better management software than these MyCloud devices, but if you get friendly with the ssh enabled console, the manual administration is remarkably similar for most tasks.
To be an effective cloud solution, you will need to do what enterprise cloud providers do: Redundant and aggregated pipes to the internet (with good peering arrangements), dedicated switch hardware behind that, with mass storage backed by redundant copies done transparently with transparent failover modes, with hot spares in the array.
That precludes the convenience of a single device, low cost solution like the MyCloud.
One can host their own VPN tunnel, over a good consumer internet pipe (like domestic fiber, if you are fortunate enough to have it), but without the peering arrangements, the traffic will be subject to throttling and other QoS related headaches imposed by the ISP-- as well, the issue of having only the single link, meaning that if the interface goes down for some reason, the storage will stop being accessible. Additionally, the lack of failover modes means that if the hosting device itself has a problem (Maybe your cat knocks it off the shelf while you are out? etc.) there are no additional mirrored clone devices to be automatically promoted to continue to serve the cloud storage, meaning the storage is down.
Basically, the 99.9999% availability guarantees provided by big cloud storage firms comes from having massively parallelized and redundant hardware at all levels that can seamlessly provide services for downed nodes. That paralellization is not something you should ever expect from a home user appliance, ever.
(For reference, a typical FAS series NetApp appliance uses redundantly cabled disk trays that use either SAS or Fibrechannel as the communication medium with the head unit, so that if the linkage chain gets damaged, a redundant path to the disks in the tray array is present, and the storage will not go down. After that, the head unit has about 5 to 10 (depending on model) gigabit ethernet ports (and PCIe slots to house additional Gigabit or fibrechannel interfaces) that can be aggregated in at least 5 different configurations for either speed or fault tolerance. They come stock with 2 fibrechannel/fiber optic interconnects for connecting to another FAS head unit of the same make and model, to function as seamless failover redundancy of the actual head unit in the event of a system failure, and should be configured on a local switched fabric with redundant paths to the head units. Each head unit also has redundant PSUs, and should be powered from redundant power rails on redundant breakers. At the RAID level, FAS series appliances use a modified version of RAID4, with double parity stripes. This means a minimum of 5 disks are required in an array, any number of arrays can be in an aggregate, and virtual volumes are hosted on such aggregates. The recommended setup is for at least 1 spare disk for each array to serve as a hot spare. The system automatically does raid scrubs to test for failing disks, and will fail a disk in the array on encountering even a single failed sector. It will the then promote the hot spare, rebuild the array automatically, and continue to serve data in single-degraded mode until the new drive is rebuilt from the remaining data and parity stripes. In comparison, a WDMyCloud (single bay) has a single gigabit ethernet port, a single sata drive in the enclosure, and a single USB3 interface on the back. One of those things is NOT like the other.
And that is the SIMPLE enterprise solution!! You should see what the clustermode OntapOS8 based solutions do! YEOWZA)
There is no “easy” solution, even for Cloud Service providers. All that stuff on the Netapp FAS series device? Needs you to know what you are doing, needs you to set it up CORRECTLY at your site, needs you to monitor it and administer it as a full time job, and needs you to plan for, and supply upgrades and replacements as required.
Basically, asking for “An easy to use, consumer level device that never fails, and never poses a single headache” is like asking for a pet unicorn. Not gonna happen. Understanding the technology, understanding that cloud storage is not magical, and understanding that you have to work for what you get, are how you make a fair compromise with yourself on your remote storage needs. “Resetting expectations” was a major job skill working for NetApp support, as the “It’s MAGIC!!” viewpoint of many executives and middle managers was so endemic, that it needed constant addressing. Reality is sadly full of gotchas, caveats, and less than nice things.