Hybrid cloud storage, first look: Synology DS916+ super-NAS
For decades, the file server was at the heart of most small company networks. For many, it was the raison d'etre for having a network to begin with. Novell (remember them?) was all about file sharing. The early NT and Windows Server builds had other features, sure, but at their core, they were file servers first.
Over the years, simple network-accessible directories gave way to more sophistication. RAID (redundant array of inexpensive disk) configurations were built so that mirroring could help preserve files if a drive failed in a server. Over time, the central file server became commoditized and appliance-ized, so that a small company could buy a box, throw some drives into it, and be up and running without a full-time IT staff.
There were some early constraints to NASs. They often were slower than file servers you could build yourself, sporting less CPU power and fewer networking capabilities. The trade-off was that they provided all the fancy file system setup and configuration. RAID configuration, even today, is not for the faint of heart.
Today, though, with the increase in performance of even low-end processors, NAS boxes are fully capable file servers, with performance on par with home built server boxes. Many of them offer one or two GigE (some, even 10GigE) wired Ethernet ports, for those who want to move huge files very quickly.
On the other hand, with more and more devices on the local business network via WiFi (which is almost always slower than a hardwired Ethernet connection), demand for network performance often doesn't strain available NAS network bandwidth.
By today's standards, built with today's components, NAS boxes make excellent small business and departmental file servers, offering all the performance of a traditional file server without the hassle, and often at a lower cost. Buying a NAS means just about anyone in the department can set up and manage the server, without needing an MCSE to just boot the thing.
But small businesses now have other solutions for storing their data, ranging from full cloud applications to storage services like Dropbox. Back in the day, when I had a sales team working for me, we had a LAN-based CRM. This was before the web, and in its very early days. SaaS applications weren't yet a thing. Everyone commuted to the office and shared access to the CRM data, which lived in a record-locked share on the file server.
Today, of course, we have cloud applications like Salesforce (and its numerous competitors), and sales folks never even need to see the inside of the home office.
You can apply this transformation to almost every server-based application. I used to run a mail server. It initially ran on a small Mac. As we grew, it moved to an NT machine. Eventually, I moved the server functions to an Exchange hosting provider, but the email lived both locally in Outlook PST files and on the server. Then I moved that to Office 365, but we still had mail files on local machines. Finally, a few years back, I moved my small business fully onto the cloud using Gmail, and no mail files at all exist locally.
In addition to traditional desktop apps moving to the cloud, many of us are also moving to mobile devices like iOS and Android-based tablets and phones. Most of that data lives not on a local server but in a dynamic hybrid of on the actual device and in a cloud service provided by the app makers. No local area network-based file server is ever touched by most iOS and Android devices.
Given all this change, is there a place in the LAN for a NAS box? For that matter, does there need to be any sort of local server on the LAN, period? Well, like almost all technology, that depends on what your needs are. I see four major areas where NAS devices fit in the small business LAN.
Bandwidth is a key consideration. I'll give you two examples to illustrate the bandwidth comparison. I have a friend who has been a dedicated backer upper of data for as long as I've known her. She is amazingly disciplined. Every Friday morning, like clockwork, she burns a CD of her Documents folder.
That's her backup strategy. Period. A year or so ago, I was curious about what happens when she fills or exceeds the CD's capacity. Would she move to a DVD? A thumb drive? Would she let me hook her up to Dropbox? After all, a CD stores about 700MB, so that's not a lot of space. Heck, a single video file would blow that away. But she doesn't have video files. She doesn't have any media. All her data ever, from nearly 20 years of computer usage, added up to 52MB.
Were she to use her meager 1.5Mbps upload speed (she has the cheapest internet she can find), it would take her nine minutes to back the whole thing up. On my 20Mbps pipe, it would take about 45 seconds to upload.
Now for the second example. Last week, I posted this 7 minute 44 second YouTube video about the maximum size you can print on various 3D printers. All told, that project took 72GB for 245 files. Sure, I shoot in multiple 4K video streams, but still, one video can take a lot of data.
On my faster pipe, using this helpful bandwidth calculator, it's clear it would take about 8 1/2 hours to upload that file to a cloud-based service -- but that's only if the cloud service supports incoming connection speeds that mirror my bandwidth. In practice, CrashPlan (the service I've been using for online backups) supports a much slower upload speed, so backups can take days and days.
As you can see, bandwidth and cloud service performance can put the brakes on a cloud-only solution for some applications. Whether its video production, security camera footage, scientific data, large libraries of photos (say, from an insurance adjuster), or any number of other media-intensive applications, cloud backup can't be the first line of backup security. You've got to back up locally.
Then, let's talk about privacy. Many people are concerned about having their data, images, and personal information online. For some, it's just a basic worry. For others, for example those who are hiding from dangerous stalkers, or those who experience brutal discrimination because of their minority identities, the need for privacy is a life and death thing.
While most cloud providers have top-tier security experts doing their best to secure your data, we've seen just how porous many of these companies, medical providers, and even governments can be. If you have your data only on your local LAN, there's still a chance of a hacker tunneling through your firewall. But there's a far-reduced chance of that happening than, say, a hacker hitting one of the big cloud providers and grabbing your data merely as part of the take.
Related to the concern of privacy is jurisdictional control. Many people (and companies, as well) would prefer their governments do not have easy access to their data. Cloud providers are often required to provide access to certain government entities (or, at least, the risk of that is there). But if your data is locked on your local LAN, again, it's up to you, your front door, and your firewall to keep everything safe. You at least know where, geographically, your data will live.
Most NASes are really general-purpose server appliances. NAS appliances have grown some new capabilities over the years. Most NAS providers have app stores, where you can download everything from a Wiki to WordPress, Apache to a VPN, and a lot more. If you want a local service (even if you still want to run your email off a local box), most modern small-business NAS boxes can do it for you.
My use has changed over the years. There was a time when I served millions of webpages a month out of a closet in my house, with a dedicated T-1 line running over the bathroom mirror. I don't do that anymore. I have to say that managed hosting providers and cloud infrastructure services like AWS and Digital Ocean make that no longer necessary. I don't miss it.
Today, I still have a lot of storage. Most of of the big capacity usage falls into three categories: video production, telemetry data from various projects, and virtual machines. Each of these areas of use generate very big files, new files are created regularly, and they're painful to send to a cloud backup service.
I use NAS boxes for most of this heavy storage. I do back all of it up to the cloud (and to multiple NAS boxes locally), because I like RAID, but I also like redundancy on different hardware. I also use my NAS boxes for smaller files, like all the spreadsheets, Word documents, and PDFs needed to just run my business activities.
So far, I've used direct-attached RAID storage for my media library. The older NAS boxes I've purchased were not as fast as USB 3.0 connections to local RAIDs. Whether that needs to remain this way as we move to faster NAS boxes and faster Ethernet remains to be seen.
In summary, I use NAS boxes as file servers. I don't use them for many other services, because most of the shared apps (email, contacts, document library, etc) are in the cloud. I use additional NAS boxes to backup both my NAS and direct-attached RAIDs. As a way of making sure I also have an offsite backup, I send a complete backup of everything to the cloud as well.
Over the past few weeks, NAS boxes have arrived here from Synology, ioSafe, QNAP, Western Digital, and TerraMaster. I also have NAS and direct-attached Drobos, which I've been using for years.
In the next month or so, I'll be unboxing the new boxes, and subjecting them to a variety of tests. Once I get a good feel for how each performs, I'll update my Battle of the Desktop NAS article with a detailed comparison of all the contenders.
The bottom line is simple: unless you store just the teensiest amount of data, you probably should have a NAS, or some other form of local backup, in addition to what you put in the cloud. Which will you choose, and how will you go about it? That's a story for another day.
You can follow my day-to-day project updates on social media. Be sure to follow me on Twitter at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, and on YouTube at YouTube.com/DavidGewirtzTV.