Nighthawk S8000 swoops down on managed switch market

Buyers of high-end consumer routers have sent manufacturers the message that aggressive styling based on next-generation fighter aircraft and science-fiction space ships helps move product. Netgear's Nighthawk series of routers was perhaps the first to fly down that corridor, and the company is now hawking a managed switch with aggressive aesthetics and a performance-oriented feature set. The Nighthawk S8000 is an eight-port managed Gigabit Ethernet switch with QoS, traffic prioritization, link aggregation, plus the angular looks and not-boring-green lights that Nighthawk products are known for.

While Netgear pitches the device to gamers and 4K streamers, the link aggregation feature could be beneficial to users of NAS or home server systems with high traffic requirements. That feature allows up to four ports on the switch to connect to the same device for 4 Gbps of combined bandwidth. According to PCPerspective, the link aggregation can also be used for failover in the event that a network cable fails. In the time it takes to watch the dubstep-infused 16-second teaser video below, 7.5GB could have passed through a four-port aggregate line, according to Netgear's claimed maximum speed of 470 MB/s.

The switch is managed with a web interface as usual, but Netgear says its GUI has been optimized for use on mobile devices like smartphones and tablets. The management features allow for three tiers of port-level traffic prioritization, as well as QoS management to ensure that gaming sessions are not derailed by things like peer-to-peer traffic. PCPerspective says the Nighthawk S8000 will cost $100 when it ships in March, and will be backed with a three-year warranty.

Comments closed
    • Gasaraki
    • 3 years ago

    Boring… Gigabit switches are a dime a dozen. Come out with a 5 port 10Gb consumer switch and then we’ll talk.

    • Chrispy_
    • 3 years ago

    Do consumers really need managed switches? The devices don’t really talk to each other enough to overload even the cheapest of unmanaged switches in a typical home. 99% of the bottleneck is the ISP or the router anyway.

    I have a NAS and run an HTPC diskless so all of its storage is a network load – and I also Stream Steam games to it from a desktop whilst the internet is being used for netflix/movie streaming to wireless devices. The cheap 5-port unmanaged switch integrated into my ISP-provided cheap router has never skipped a beat and I’d be amazed if management was required for a gigabit switch with as few as 8 ports. If you have high QoS and prioritization issues on your LAN and not your WAN connection, then you really should be looking at a proper layer-3 switch. If gigabit isn’t fast enough for you, then you’re probably not a consumer.

      • Waco
      • 3 years ago

      I wouldn’t mind a managed switch, but you couldn’t pay me (at this point) to buy a gigabit switch with 2.5g and 5g stuff on the horizon for the consumer market.

      That, and my need for a managed (LACP capable) switch diminishes heavily as soon as those faster techs become available…

    • LoneWolf15
    • 3 years ago

    As “dull” as they might look, good switches are rectangular for a reason: So you can put things on top of them easily, and make good use of space.

    I will never be buying a switch meant to look “edgy” and “artisan”. I can get square managed gig switches from Netgear and HP with the same number of ports in a similar price range.

      • chuckula
      • 3 years ago

      Rack mountable: It’s not just a good idea, it’s the LAW.

    • kurazarrh
    • 3 years ago

    “… in the event that a network cable fails.”

    I install and manage networks small- to medium-size businesses for a living. I can count on one hand the number of times a network cable has failed, and 100% of those failures were because the client location’s employees had run long patch cables from wall jacks to their computers and routinely walked on and frayed the cables, or in one case, they installed the cable too close to some equipment that slowly melted the Cat5 shielding. What are people doing to cables that they need aggregate links in case the cables fail? It ought to be common knowledge that the NIC/port is thousands of times more likely to fail than the actual cable, which is hardly anything special (8 shielded copper wires).

      • cygnus1
      • 3 years ago

      Agreed, a redundant failover between ports is almost always going to happen because of a failed port or misconfiguration of a port.

      • the
      • 3 years ago

      It is less completely failure and more a cable becoming loose that isn’t obvious. Physical cable is fine but gets move just enough in a port to terminate a connection. Reseating resolves the issue but you’d still want fail over in this case.

        • Brainsan
        • 3 years ago

        I’ve got around 500 computers at work. Twice I’ve had gorillas or pro wrestlers (I was previously unaware that we hired either) shear connectors off of network cards while “reseating” cables.

        Of course they were reported as “we can’t plug in the network cable”, not “I’m an idiot that should never touch a computer”.

      • derFunkenstein
      • 3 years ago

      The only time I had a cable fail is when I tripped over it because of a weird desk/port placement decision (not mine, someone else’s office). Ripped the connector off the cable and miraculously the machine wasn’t damage. My guess is that the wires were already frayed, because that was freakish.

      • curtisb
      • 3 years ago

      While I agree with you, it’s becoming a bigger issue in enterprises with more PoE gear. I’m not talking VOIP phones that are fairly low power, but more Wi-Fi AP’s. We were talking with one of our engineers at the vendor we use for network gear and they said they’ve found that over time (years, at that) the cable quality becomes degraded. It’s not so much in the run as it is in the bundle of cables at the rack. This is because of the additional heat created by pushing power over them, and the bundles not allowing the cables to cool off. The next time we replace our AP runs we’re going to separate them from the rest of the cable bundle.

      It’s still going to take years, but I completely agree that LAG’s are more about throughput than redundancy due to cabling. The redundancy aspect comes in from NIC or switch failure, and a switch failure isn’t going to be covered by a normal LAG.

    • DreadCthulhu
    • 3 years ago

    Seems strange to give it a stealth design; you [i<]want[/i<] a router to pick and send radio waves, after all.

      • Vaughn
      • 3 years ago

      this is a switch not a router it has no Wireless radio’s!

        • nico1982
        • 3 years ago

        Try again 😛

        • curtisb
        • 3 years ago

        A “router” is not automatically wireless. And technically a fully-managed L3 switch is a router. 🙂

          • Vaughn
          • 3 years ago

          Any consumer level router bought in the last 5 years is automatically wireless they don’t ship them with the radio’s turned turned off.

    • Waco
    • 3 years ago

    Four gigabit (112.5 MB/s) links…Gives 470 MB/s? Math must be hard for them.

      • chµck
      • 3 years ago

      [quote<]That feature allows up to four ports on the switch to connect to the same device for 4 Gbps of combined bandwidth.[/quote<] It's an aggregate line, so why not? Also, isn't 1 Gb = 125MB?

        • Waco
        • 3 years ago

        Maximum TCP throughput in perfect conditions should be 116 MB/s without any network trickery (like jumbo frames): [url<]https://kb.netapp.com/support/s/article/what-is-the-theoretical-maximum-throughput-of-a-gigabit-ethernet-interface?language=en_US[/url<] I'm not sure if LACP adds any additional protocol overhead. I just use 112.5 as an average gigabit speed since you'll very rarely see anything exceeding that. 470 MB/s would imply they're using jumbo frames in a perfect lab test. Granted, it's not much different than the 450 MB/s you could potentially see with perfect scaling, but even that's pretty damn hard to achieve.

          • cygnus1
          • 3 years ago

          It’s an absurd suggestion that anyone will ever see even more than 1gb out of it. Very few consumer client devices are ever going to have multiple NICs. And even if they do, LACP does not “channel bond” how most people assume, a single TCP stream can only go out 1 port which means they’d need to ensure software on both ends that can multiplex across multiple streams to saturate multiple links.

            • hans
            • 3 years ago

            Balance-round Robin mode will stripe across the bond slaves, giving the full aggregate speed between just two hosts. Mode 0 in the Linux bonding driver. Still requires a port channel on the switch.

            • cygnus1
            • 3 years ago

            And what’s the plan with an 8 port switch? Connect two Linux servers with quad NICs to the switch and let them talk to each other like that? Might as well just do direct connections between the hosts at that point, you’re out of switch ports and aren’t going to talk to any other systems anyway.

          • Laykun
          • 3 years ago

          I think that’s why they used the word “maximum”. You’re not expected to see these results, but at it’s most practically efficient, you can achieve 470MB/s peak.

            • Waco
            • 3 years ago

            I guess. This is where marketing tends to jade people though, because if they’d just said 400 MB/s, they’d have nobody doubting their methods.

      • nico1982
      • 3 years ago

      It is the press material of a consumer grade $100 router…

Pin It on Pinterest

Share This