ASUSTOR Unveils AS-U2.5G2 Portable 2.5GbE Adapter

[btarunr]

ASUSTOR, the NAS-focused brand of ASUS, unveiled the AS-U2.5G2, a pocket-size external network adapter for 2.5 Gbps Ethernet. Built into a premium casing made of aluminium, the adapter uses a Realtek RTL8156B 2.5 GbE PHY, which is a single-chip solution designed for USB 3.0, without the need for a USB-to-PCIe bridge chip. A single USB 3.2 Gen 1 connection handles both power and host-connectivity. The adapter comes with a convenient USB type-C connector, and a type-C to type-A dongle is included. It measures 16.4 (H) x 23.4 (W) x 195.2 (D) mm, and weighs just 33 g. Since it's being sold standalone, it's compatible with not just ASUSTOR NAS products, but also any Windows 10 PC. The company didn't reveal pricing.



View at TechPowerUp Main Site

 

[Valantar]

"Built into a premium casing made of aluminium", lol. As is every $10-and-up USB-to-Ethernet adapter these days. Nothing premium about a tube of extruded aluminium, even if it's nicer than plastic.

Hopefully this comes in at $30 or below, it's good to see some competition in this space.

 

[TheLostSwede]

It retails for about $40 in Taiwan.

 

[Aht0s]

Why not go to 10Gb instead.
Would converting from USB 3.0/1 to USB-C to TCP/IP add overhead?
I can't think of a situation that I would need it in a NAS environment.

 

[TheLostSwede]

Why not go to 10Gb instead.
Would converting from USB 3.0/1 to USB-C to TCP/IP add overhead?
I can't think of a situation that I would need it in a NAS environment.

Uhm, for several reasons, the first being 10Gbps isn't possible over USB 3.0, as the USB interface is limited to around 4.8Gbps based on a best case scenario.
I think you're very confused with regards how to this thing works. Any and all Ethernet interfaces would have a TCP/IP overhead, the physical interface to the computer has nothing to do with that.
2.5Gbps is the "new" standard which will (hopefully not too) slowly replace 1Gbps Ethernet. Mostly due to it not costing a lot more per chip and the fact that it works great over Cat5e.
10Gbps requires Cat6A for longer runs, which means new cables. Sure, it's a lot faster, but you'll also spend a lot more on switches, not even taking cabling into account.
QNAP has a $100 5-port 2.5Gbps switch already.

 

[repman244]

Why not go to 10Gb instead.
Would converting from USB 3.0/1 to USB-C to TCP/IP add overhead?
I can't think of a situation that I would need it in a NAS environment.

Because 10G Base-T is not efficient as let's say sfp. You are probably looking at ~10w of power, but yes I do agree 10G would be way better but I guess the industry wants to milk it slowly and firat introduce 2.5G, 5G and then finally go to 10G.

 

[Aht0s]

Uhm, for several reasons, the first being 10Gbps isn't possible over USB 3.0, as the USB interface is limited to around 4.8Gbps based on a best case scenario.
I think you're very confused with regards how to this thing works. Any and all Ethernet interfaces would have a TCP/IP overhead, the physical interface to the computer has nothing to do with that.
2.5Gbps is the "new" standard which will (hopefully not too) slowly replace 1Gbps Ethernet. Mostly due to it not costing a lot more per chip and the fact that it works great over Cat5e.
10Gbps requires Cat6A for longer runs, which means new cables. Sure, it's a lot faster, but you'll also spend a lot more on switches, not even taking cabling into account.
QNAP has a $100 5-port 2.5Gbps switch already.

Gotcha, thanks for the clear up. 3.1 could though no? Adapters/cables are reasonably cheap.

Because 10G Base-T is not efficient as let's say sfp. You are probably looking at ~10w of power, but yes I do agree 10G would be way better but I guess the industry wants to milk it slowly and firat introduce 2.5G, 5G and then finally go to 10G.

Thanks, I didn't think of the efficiency part. They are definitely milking it... I understand is overkill for home network, but it feels like we should have access to this kind of speed by now...

 

[TheLostSwede]

Gotcha, thanks for the clear up. 3.1 could though no? Adapters/cables are reasonably cheap.

Nope, 3.1 still doesn't get you 10Gbps of usable bandwidth, at least not from anything I've seen, as it's still using 8b/10b encoding.
You might get close to it if you go with USB 3.1 Gen 2/USB 3.2 Gen 2x1 (I still can't get over these stupid naming schemes) as that's using 128b/132b encoding, so you get closer to the actual 10Gbps speed of the interface.
That said, today there's not a single 10Gbps Ethernet controller that draws as little as 4.5W, so you'd also need a port that supports USB PD and then things get super complex, as not all USB ports supports USB PD, even USB-C ports.
In theory, you could make one, but there would be so many caveats that it doesn't make sense.
You can get 5Gbps Ethernet dongles though, as Aquantia made a 5Gbps Ethernet to USB controller before they got bought by Marvell. It does require USB 3.1 at a minimum though.

 

[Aht0s]

Nope, 3.1 still doesn't get you 10Gbps of usable bandwidth, at least not from anything I've seen, as it's still using 8b/10b encoding.
You might get close to it if you go with USB 3.1 Gen 2/USB 3.2 Gen 2x1 (I still can't get over these stupid naming schemes) as that's using 128b/132b encoding, so you get closer to the actual 10Gbps speed of the interface.
That said, today there's not a single 10Gbps Ethernet controller that draws as little as 4.5W, so you'd also need a port that supports USB PD and then things get super complex, as not all USB ports supports USB PD, even USB-C ports.
In theory, you could make one, but there would be so many caveats that it doesn't make sense.
You can get 5Gbps Ethernet dongles though, as Aquantia made a 5Gbps Ethernet to USB controller before they got bought by Marvell. It does require USB 3.1 at a minimum though.

cool, thank you for the explanation.

 

[Tartaros]

Probably we won't see 10Gbps being pushed to mainstream for at least a decade or longer. Unless there is some sort of big jump in wan technology which is cheap and easy to implement, which never is.

Gigabit ethernet has been around for 15-20 years and recently the mainstream users started to use it to it's potential.