Over the years, I was constantly improving my way to handle and backup data. During the end of last year, I was thinking of an architecture that suits best to my current needs and demands in terms of:
- security of data: it shouldn’t be too easy to hack my storage
- easy backups of my data: I don’t want to manually backup my data anymore
- availability of data: I want to be able to access my data even if my local internet doesn’t work or there is a local power failure
- accessibility of data: I want to access my data from everywhere and with reasonable access time
- reasonable costs of the solution: since I’m not running a business but I’m rather a private person hosting private and mostly non-sensitive data, the solution should fit to the budget of a private person as well
- sustainability of the solution: I want the storage solution I use to be scalable, so it can grow with my data
- compatibility of the solution: I want to use standard components in my solution wherever possible to make sure the solution is compatible to most devices and services I use
In this post I want to introduce what I came up with.
I decided to distinguish between three different kinds of layers in my data architecture:
- Device Layer
- Cloud Layer
- Local HDD Layer
…and it looks like this:
Device Layer
The device layer represents all devices I use that need to access my data.
Most of my devices are mobile devices (except my TV and my Xbox One), so they need to have access to data from everywhere using the Cloud Layer. My TV and my Xbox One only need to access the data from home, but whereas OneDrive is integrated nicely into Xbox One, many smart TVs are still missing a professional cloud support and therefore I need to refer to my Local HDD Layer.
Cloud Layer
The Cloud Layer acts as single source of truth for my data.
Most devices I use shall preferably save data directly into the Cloud Layer. From here, my data will be synced to most other devices and also to the Local HDD Layer. Of course this means that also the folder structures and changes I apply to it will therefore always be transferred to all other devices as well and I don’t have to do new folder structures manually on all devices like in earlier times.
I decided to go with Microsoft OneDrive as a cloud service since it is compatible to most devices I use, the pricing is really reasonable and it supports a lot of useful features. I especially like how OneDrive is integrated into the Microsoft ecosystem and the services/apps that are provided by Microsoft, e. g.:
- OneDrive is the standard cloud of Microsoft and therefore well integrated into Windows 10
- I can use Office Online to edit all my documents from anywhere and any device
- I can show my pictures on my Xbox One or the Xbox One of any of my friends
- I can use the OneDrive App so that pictures I take with my smartphone will be uploaded to OneDrive automatically and I don’t have to transfer pictures manually anymore
- I can share pictures of trips or events right away through the cloud with my friends without special efforts (even albums for sharing are created automatically now with OneDrive)
And yes, I believe in the security of cloud storages if you use safe passwords and two-step-verification. Of course, your data is never entirely secure, but even if I use my own private cloud it can be hacked and the performance will even be worse. Also, using a cloud can be your last way to access your data via smartphone if you experience a local power failure.
Local HDD Layer
The Local HDD Layer represents all local copies I have from all my data which is saved in the cloud.
I’m currently using a Synology single-bay drive which I use as a „cloud gate“. The cloud syncs directly to the Synology DS115j and makes all cloud data available offline in my home network (see instruction how to sync your cloud with your Synology drive). Therefore, I do not need to equip my private laptop with a huge HDD to access all data I have stored into the cloud: I simply use my Synology to access the data or even to copy large amounts of new data into the cloud. My ThinkPad only has a 128 GB SSD and therefore feels like a lean thin client. Also, if my laptop gets stolen, the offender doesn’t get all cloud data on the local device. I do sync my documents to the laptop, but neither pictures nor videos or software backups are synced directly to my laptop because they’re simply taking to much space on my SSD drive and I don’t need to carry these types of data with me. Since Synology not only supports OneDrive but also Amazon Cloud Drive, I can decide to use Amazon Cloud Drive additionally at a later stage of time, if the 1 TB of space OneDrive offers to me is not enough anymore (Amazon Cloud Drive offers unlimited cloud space at decent prices).
Also, I am using my Synology as a media station in my home network. All DLNA supporting devices in my home network may therefore access my data on the Synology even if there is no internet connection available. My TV and even my DECT FRITZ!Fon are connected to the Synology instead of the cloud, since there are no (suitable) apps for OneDrive (or any other cloud service) available to connect to the cloud directly .
Additionally I’m using a QNAP dual-bay drive running in RAID 1 that automatically receives incremental backups of my data from the Synology drive once a week (see instruction how to set up rsync backups between Synology and QNAP drives). The QNAP is therefore not directly synced to the cloud and also the backups it holds reach back 24 months. This means that I can undo all changes that have been applied to the cloud data within the last 24 months and I have a backup of all my cloud data on two local HDDs. If one HDD breaks down, I usually replace both at a time. And why am I using a QNAP as backup dual-bay? Because I simply have it as a left-over of my previous architecture. I will probably replace it with a Synology multi-bay by time if it doesn’t work anymore.
Costs and conclusion
Hoping that my thoughts will be useful for some of you as well, I can however recommend my set-up to anyone who has similar demands like me. My whole fixed costs are less than 800 EUR for my NAS devices and variable costs are limited to my OneDrive plan (plus eventual replacement of HDDs for maintenance issues).
Great article.
Have you thought about a third step for your backup process. You have both local HDD Devices (Synoloby and QNAP) in the same power grid. What will happen, when you get a lightning strike in your home or flat?
I save the high prio datas additionally manually on a seperatly external USB SSD Device.
I store this device outside of any electricity environments.
Hi Carlos! Thanks for your feedback! I have indeed considered the points you brought up in your comment.
Firstly, the QNAP is located on another floor in my home, so if a fire breaks out on my floor, the QNAP will not be affected.
Secondly, in case of any lightning strike, my data is still save in the cloud 🙂
Lastly, I am still thinking about one more Synology to be put at a friends home which I will then sync via Internet with my other Synology as a second backup.
I don’t like to backup manually on external devices anymore because usually if anything happens you didn’t do the last backup on exactly the day before 🙂