When The Spark Fades…

My decision to buy my first “proper” NAS box, back in 2012, was  really driven by the mixed experiences I’d had with an Apple Time Capsule. In principle the ATC is a fantastic idea; in practice Apple designed it down to a price, chose poor quality components and mine – like so many of the first generation units – burnt out it’s PSU after about 2 years [just outside warranty].

I replaced that with a QNAP TS-459 [which has been brilliant], and did really well until the first set of HDDs I’d purchased [Western Digital Greens, 3Tb each] began to fail. A hurried purchase of a larger/newer/better TS-670 followed, with this crammed out with no less than six Western Digital Reds, 6Tb each, configured in to RAID 6. The older 459, meanwhile, took a brace of 4 x 6Tb Reds, configured to RAID 5, and now exists primarily to be a “local” backup to the 670. [I have the machines configured to run a 7-generation backup at 2am every morning – works brilliantly].

Except.

Not even this can cope with a power outage. Enter one of these:-

It’s a “UPS” – an “Uninterruptible Power Supply”, basically a battery, a mains conditioner and an inverter, all in a neat little box that measures about 25cm x 25cm x 10cm and has the ability to power up to 4 devices for a reasonable period of time in the event of a power outage…  From experience, it’s uncommon to get a power failure that lasts longer than a couple of minutes [usually it’s just the electricity company doing some sneaky maintenance in the middle of the night]. In other words, this should be more than enough to keep both NAS boxes nice and safe, even through a power cut.

Just to be on the safe side, this UPS also has a cable which I’ve connected to my NAS. If the power does fail, the UPS will alert the NAS, which have the ability to shut themselves down smoothly, before the batteries fail.

Not bad for £60 + cables; let’s hope that now I’ve got this, I won’t actually need it…

Lies, Damned Lies – and Claims from Powerline Adapter Manufacturers!

Among undoubted charms, one of the few little challenges presented to me by my current home is the fact that I have no easy way to run network cables around. All the walls are brick; the floors are concrete, there’s nowhere to conceal cable runs. As a result, I’ve invested in Powerline adapters, little units that come in pairs and which you can plug in to a couple of mains sockets and which, if you’re lucky, can route Ethernet traffic around your property using the heavy gauge mains cable as a transmission medium.  This kinda works…

Back in 2015 I purchased a pair of adapters with a claimed throughput of 1.2 Gigabits per second. Since they were massively faster than the 500Mb/s units they replaced, I was quite happy with them, although I noted at the time that they didn’t quite deliver as claimed…

This evening I’ve just installed and tested an upgraded pair of adapters, this time boasting a claimed 2000Mb/s [2 Gigabits] of throughput. So fast, in fact, that I should expect them to be at least as fast as a direct network connection, right? Ho, ho, ho…

I ended up running 30 tests across 3 different scenarios:-

1. Direct – Laptop with SSD to Gigabit Switch to QNAP TS670 NAS
2. Legacy via 2 x AV1200s over approx 10m of mains ring
3.  New via 2 x AV2000s over approx 10m of mains ring

For the Powerline Tests I performed power-recycle resets and then conducted 3 full test runs to let the adapters “train” themselves before capturing data. Each test was conducted as pairs: one test was for a large single file [approx 9Gb]; the other was for approx 250 files in a folder structure with the total size approx 9Gb].

For each of the above configurations and the two tests, I ran each test 5 times. I then averaged down the 10 results for each configuration to a single value and report the results as bytes-per-second not bits-per-second…

1. Direct Connection – 153,294,088.732 bytes per second
2. Powerline AV1200s – 16,193,844.940 bytes per second
3. Powerline AV2000s – 27,324,602.086 bytes per second

In other words, despite claiming to be twice as fast as conventional Gigabit Ethernet, the AV2000 units actually only returned – as you can see from the above data – about 17.82% of the performance of a direct connection. From equipment claiming to have twice the performance of a direct connection…

Sum-Up: Powerline Ethernet Manufacturers: full of sh#t!

[ Oh, and in case you’re wondering – the 1200 units were perfectly capable of letting me watch streamed 4K content from YouTube, Netflix and Amazon Prime, delivered down a BT Infinity 75Mb/s FTTC service – so the 2000s will be more than sufficient]. They’re good, just nowhere near as good as claimed… ]