Happy with my EA6900

[htismaqe]

Well, so far, so good on the EA6900.

Entering initial configuration was kind of a pain because the Linksys GUI wants to force you to use their Cloud service. I much prefer configuring offline and turning up fully secure and on my own terms.

Once I actually got into the router, setup was a snap. Not anymore difficult or easy than Netgear Genie but A LOT faster. First, not every settings change requires a reboot of the router, as it seems every Netgear device requires. Additionally, even when the Linksys does need to reset or reboot, it takes like 10 seconds instead of over a minute each time.

I'll be putting it into production later today and I'll be able to provide some side-by-side comparison on wireless range and performance then.

 

[htismaqe]

OK, I've only played with the antennas a little bit but all in all, it doesn't quite have the range as the Netgear on 5Ghz - maybe 6-8dB weaker at my farthest location is all, though. 2.4Ghz is almost identical to the R7000 and in my family room 35-feet from the router, the EA6900 is actually a bit stronger. The antennas on the EA6900 aren't very easy to work with though, weird little things.

Despite the difference in range, I really haven't found a difference in throughput at all. Even in my office (the farthest location I mentioned above) read/write speeds to my NAS are within a few MB of the R7000.

I did find one fairly significant difference in the wireless "performance" though - I have 4 private SSIDs in my house, the 2 on the router (2.4Ghz and 5Ghz) and two external 2.4Ghz APs. While I was on the R7000, switching SSIDs on a device could take anywhere from 5-15 seconds. In some cases, I'd have to reconnect network drives because it took long enough that they would time out.

On the Linksys, when I switch networks, it reconnects immediately. There's not even a 1 second delay. It's instantaneous.

And I have to say, two things that I like better in the EA6900 - DHCP reservations and Guest Wifi.

The Guest Wifi is setup like I'm used to seeing in commercial and enterprise networks - open SSID, no PSK, and splash page for access. Guests are sandboxed on their own IP subnet and completely isolated from the interior LAN. Big difference from the way Netgear implements it.

DHCP reservations also work more like I'm used to in business-class routers because the reservations have to actually be IN your DHCP scope. On the R7000 (and several other SOHO routers I've used over the year) you could set a scope of say 192.168.1.10-25 for your dynamic clients and then do reservations on 30, 31, 32, etc. You can't do that on the EA6900. Even though there's less data entry on the 2nd method, the first method for me is a lot cleaner and easier to use after initial configuration.

Finally, speaking of initial configuration - despite the fact that there was almost 3x more data entry for me to port my DHCP reservations manually to the EA6900 from the R7000, it took me less than HALF the time. With the R7000, you have to enter a 2nd screen to add individual entries, so you spend almost as much time navigating through the pages as you do entering reservations. On the EA6900, you can enter them all at once, edit them as many times as you like, and then apply them all right on one page. Not as fast as loading them through a shell script like on Rmerlin firmware or DD-WRT but by far the best GUI implementation I've ever seen.

Overall, I'm liking this router a lot, so I'd like to thank Mr. Higgins for the suggestion. Here's to hoping it stays stable, which ultimately was the reason I abandoned the R7000.

 

[htismaqe]

A couple of other quick things...

The EA6900 appears to be able to use per-user authorization for access to the USB drive, which is one of the main reasons I never used Readyshare on either my WNDR3700 or the R7000. Netgear only allows use of the admin account meaning everybody on your network has access to all of the files on your drive.

I like the fact that I can use HTTPS for administering the router. Not an option in Netgear stock firmware.

Finally, I experienced a strange issue with the R7000 where uploading large files to my NAS via AFP (from my iMac) would consume 100% of my wireless bandwidth, making it impossible to multi-task. I had a ticket open with Netgear for over a month and they never did figure it out. The EA6900 works great with AFP so far.

 

[sfx2000]

A couple of other quick things...

The EA6900 appears to be able to use per-user authorization for access to the USB drive, which is one of the main reasons I never used Readyshare on either my WNDR3700 or the R7000. Netgear only allows use of the admin account meaning everybody on your network has access to all of the files on your drive.

I would be very, very careful with Samba shares... for linksys, it's share level access for some implementations, not user level access...

Not trying to scare you, but just be aware... I would nmap the device in your preferred connection to be sure...

Below is a quickscan against WRT1900AC as default;

|_msrpc-enum: NT_STATUS_BAD_NETWORK_NAME
| nbstat: NetBIOS name: LINKSYS11008, NetBIOS user: <unknown>, NetBIOS MAC: <unknown> (unknown)
| Names:
|   LINKSYS11008<00>     Flags: <unique><active>
|   LINKSYS11008<03>     Flags: <unique><active>
|   LINKSYS11008<20>     Flags: <unique><active>
|   \x01\x02__MSBROWSE__\x02<01>  Flags: <group><active>
|   WORKGROUP<1d>        Flags: <unique><active>
|   WORKGROUP<1e>        Flags: <group><active>
|_  WORKGROUP<00>        Flags: <group><active>

...

| smb-mbenum:
|   DFS Root
|     LINKSYS11008  0.0  Samba 3.0.28a
|   Master Browser
|     LINKSYS11008  0.0  Samba 3.0.28a
|   Print server
|     LINKSYS11008  0.0  Samba 3.0.28a
|   Server
|     LINKSYS11008  0.0  Samba 3.0.28a
|   Server service
|     LINKSYS11008  0.0  Samba 3.0.28a
|   Unix server
|     LINKSYS11008  0.0  Samba 3.0.28a
|   Windows NT/2000/XP/2003 server
|     LINKSYS11008  0.0  Samba 3.0.28a
|   Workstation
|_    LINKSYS11008  0.0  Samba 3.0.28a
| smb-os-discovery:
|   OS: Unix (Samba 3.0.28a)
|   NetBIOS computer name:
|   Workgroup: WORKGROUP
|_  System time: 2014-04-26T01:52:24+00:00
[B]| smb-security-mode:
|   Account that was used for smb scripts: guest
|   Share-level authentication (dangerous)[/B]
|   SMB Security: Challenge/response passwords supported
|_  Message signing disabled (dangerous, but default)
|_smbv2-enabled: Server doesn't support SMBv2 protocol

 

[htismaqe]

Yeah, it's not an issue, just an observation.

1) I don't use USB on my routers because I have a NAS.

2) I don't use SMB between my iMac and my NAS because SMB is pretty broken in Mavericks.

I was merely pointing out that at least the Linksys appears to have some level of differentiation between different users, even if you're assigning it at the share level.

With the Netgear stock FW, everybody is "admin", accessing file sharing uses the same admin password as the admin webUI, and thus all folders are open to anybody that logs in. It's one of the biggest reasons I never used it, dating all the way back to my first WNDR3700.

 

[htismaqe]

2 more things:

1) For whatever reason, the EA6900 fails a GRC Shields Up! test because port 445 is closed instead of stealth. I need to look into it further but since that port is used by MS and is commonly attacked, it's a little annoying.

2) One thing I read (and was warned about) with the EA6900 was that it was prone to overheating. As of this morning, it's barely running warm, let alone hot. I'm going to say it's actually running just a tad cooler than the R7000. I did do something however, to make sure it never overheats. You can read more here:

http://forums.smallnetbuilder.com/showthread.php?t=17351

 

[htismaqe]

I really like the GUI Device List in the Linksys compared to Netgear.

The biggest advantage for the Linksys is the ability to edit device names and have them stick. The Netgear GUI allows you to enter custom device names in a few places (e.g DHCP reservations) but they're not globally significant so when you go to view "attached devices", half of your devices have 2 dashes and a MAC address or names like "android-123456". On the Linksys, you can override those in the GUI Device List - much cleaner and easier to use.

I will give the Netgear one advantage though - it separates devices by wired and wireless. So far from what I can tell, Linksys lumps all connected devices together on one page, although it does separate LAN from Guest, which the Netgear does not.

 

[htismaqe]

Here's something interesting...

I have 30Mbit ethernet over fiber for my Internet connection, nothing special.

When connected from my iMac over 5Ghz to the R7000, ping times from my machine to the R7000 were about 1ms and ping times to my ISP's primary name server were about 6ms.

When connected over 5Ghz to the EA6900, ping times to my ISP's name server are only 4.5ms. Ping times to the router itself are in the 1-1.5ms range.

In fact, I'm seeing anywhere from 10-25% lower roundtrip response times in general with the EA6900.

 

[sfx2000]

Here's something interesting...

I have 30Mbit ethernet over fiber for my Internet connection, nothing special.

Hopefully it's symmetric - based on our offline conversation, I wouldn't expect any less

When connected from my iMac over 5Ghz to the R7000, ping times from my machine to the R7000 were about 1ms and ping times to my ISP's primary name server were about 6ms.

When connected over 5Ghz to the EA6900, ping times to my ISP's name server are only 4.5ms. Ping times to the router itself are in the 1-1.5ms range.

In fact, I'm seeing anywhere from 10-25% lower roundtrip response times in general with the EA6900.

Not bad in any event...

I'm too far away from the local ATT Central Office to get decent DSL, so I've had to bank on cable, which they're pretty good, and unlike some cableCo's, CoxHSI seems to be competent, and their customer care is very good.

I agree with you - Linksys does push SmartWiFi pretty hard, but there is a link on the initial setup screen to bypass the cloud solution.

Sounds like it's a close cousin of the WRT1900ac - main different being Broadcom vs. Marvell based - 6 of one, half-dozen of the other.

Thanks for the info.

sfx

 

[sfx2000]

Yeah, it's not an issue, just an observation.

1) I don't use USB on my routers because I have a NAS.

2) I don't use SMB between my iMac and my NAS because SMB is pretty broken in Mavericks.

Yep, SMB on Mavericks is due to the Samba team adopting GPLv3, anyways, there is an app that will switch Apple's SMB stack from SMB2 to SMB1, where it does work better with Samba

http://www.macparc.ch/apps/SMBconf/

I was merely pointing out that at least the Linksys appears to have some level of differentiation between different users, even if you're assigning it at the share level.

True - for most folks, it's good enough... but not the most secure way to setup a samba share...

With the Netgear stock FW, everybody is "admin", accessing file sharing uses the same admin password as the admin webUI, and thus all folders are open to anybody that logs in. It's one of the biggest reasons I never used it, dating all the way back to my first WNDR3700.

Yikes!

sfx

 

[htismaqe]

I agree with you - Linksys does push SmartWiFi pretty hard, but there is a link on the initial setup screen to bypass the cloud solution.

Yes, it ended up being easy to at least get to a login prompt. The problem at that point was that nowhere in the documentation did it list the GUI password. I had to get out my iPad, connect to 3G, and lookup the admin password in order to get into the thing.

Yep, SMB on Mavericks is due to the Samba team adopting GPLv3, anyways, there is an app that will switch Apple's SMB stack from SMB2 to SMB1, where it does work better with Samba

I had read somewhere that you could force SMB1 by using "cifs://" instead of "smb://" but I never could get it to work. Thanks for that tip, in case a need to use it in the future. Right now, AFP is working great so I'm happier than I've been in over a month.

 

[sfx2000]

Yes, it ended up being easy to at least get to a login prompt. The problem at that point was that nowhere in the documentation did it list the GUI password. I had to get out my iPad, connect to 3G, and lookup the admin password in order to get into the thing.

Interesting - there wasn't an insert card blown into the box that had the admin PW? Also, at least on the 1900ac, it's printed on the bottom of the router on the same sticker with the MAC address and other goodies specific to the device...

I had read somewhere that you could force SMB1 by using "cifs://" instead of "smb://" but I never could get it to work. Thanks for that tip, in case a need to use it in the future. Right now, AFP is working great so I'm happier than I've been in over a month.

Yah, I shared that advice about trying CIFS for the mountl - but the switcher app forces the issue if you need SMB1... there's been much grumbling about the SMB issue...

sfx

 

[htismaqe]

Interesting - there wasn't an insert card blown into the box that had the admin PW? Also, at least on the 1900ac, it's printed on the bottom of the router on the same sticker with the MAC address and other goodies specific to the device...

Nope. The insert only lists the wifi SSID and PSK. They really expect you to connect it to the Internet, connect via wifi, and set it all up like that. I didn't look real close but the plaque on the bottom only has the MAC and S/N IIRC.

 

[Nexus]

Glad to see you found a router you like. I've decided to hold off on my search at the moment. The new fiber install takes place in August, so I'll wait until then and hope that some of the higher end AC models drop in price a bit by then? I'll definitely keep this particular Linksys model in mind because I like some of the features you listed here.

 

[htismaqe]

Yeah, it's probably worth waiting.

All of these AC1900 routers seem to have at least some issues it seems.

I've been over at the Linksys forums tonight and there's a ton of issues with the EA6900 crashes and losing connectivity.

However after reading most of them, I honestly think it's a thermal issue and putting some air under the router will help immensely. I did find this post, which I thought was interesting:

At 5:40PM MST, my newly unboxed AC6900 went online using default settings accept for using 192.168.0.1 for the router IP address, Mixed mode on both 2.4Ghz and 5Ghz, manual channel 11 and 48 perspectively. Reserved IP addresses for all devices accpet for 4 devices, 2 NAS, Airport Express and one printer. ATTMicrocell in the DMZ. Xbox set in Media Prioritiztion and connected to a wireless AC bridge connected to the 6900s 5ghz radio. Apple Airport Express set up for static and bridged to the 2.4ghz radio.

SB 6180(ISP Modem)>EA6900>24pt Gb network switch<>all other wired devices connected to switch.
Mac Book Pro, Mac Mini, 2 iPhone 4S, Ipad and DIR-865L (xbox) bridged and Airport Express connected wirelessly to router.

FW Loaded: 1.1.42.158863
Room temp was 80-85F. Afternoon sun heats up the office.
After router was onlie for 1 hour: Operating Temp. 32 to 104°F (0 to 40°C) listed in User Manual.
Top center section of router temp was 105-109F.
Bottom center section towards back of router temp was 125-130F.
NO external cooling...yet.
Measured using laster temp gun.

Will let run for 24 hours then check and try updated FW maybe.

 

[htismaqe]

I decided to put it through one last test today - archiving my Time Machine to NAS (not to USB on the router, a standalone NAS).

Unfortunately, when transferring that much data, the EA6900 just isn't going to cut it for me.

With small transactions, the EA6900 compares favorably to the R7000 in performance. Evidently the difference in signal strength isn't enough to affect file transfers over short periods of time.

However, with a 125GB file, the difference is more than notable.

Connected to a 2.4Ghz Netgear WGR614v10 in AP mode, estimated transfer time is just over 16 hours.

Connected to the 5Ghz on the R7000, estimated transfer time was just under 5.

Connected to the 5Ghz on the EA6900, estimated transfer time started at 6 hours and eventually ballooned to over 18. I started monitoring the wireless link and noticed that my Tx rate was moving all over the place, anywhere from 75Mbps all the way down to 18. Like I said, not noticeably really when doing web surfing or other stuff but really noticeable in this instance.

On a side note, even with 1-inch of clearance under the router, I'm noticing it's getting really hot now. A lot of reports in the Linksys forums of instability after 3 or 4 days.

Looks like I'll be trying the AC68U next.

 

[htismaqe]

WOOHOO!

After more messing, I feel much better.

I tried out a few different 5Ghz channels. After setting the channel to 153, 40Mhz width, and setting operating mode to N-only, I've gotten things to settle down CONSIDERABLY.

Transmit rate on my Mac is now fairly stable at 108Mbps and RSSI improved 6-8dB.

Current estimated transfer time for Time Machine archive is 6 hours. VERY acceptable for my purposes. I really like quite a few things about the Linksys so getting this figured out was a big relief.

 

[htismaqe]

So the issue appears to be the wireless channel.

With both the WNDR3700 and R7000, I found the best range and performance on channel 161.

Setting the EA6900 to either 161 or 165 causes the connection to become unstable garbage.

On channel 153, it is rock solid. Total estimated transfer time has dropped to 5 hours - just marginally worse than the R7000.

 

[L&LD]

So the issue appears to be the wireless channel.

With both the WNDR3700 and R7000, I found the best range and performance on channel 161.

Setting the EA6900 to either 161 or 165 causes the connection to become unstable garbage.

On channel 153, it is rock solid. Total estimated transfer time has dropped to 5 hours - just marginally worse than the R7000.

If you haven't tried the lower channels on the 5GHz band, you might want to.

When I tested the lower channels on my RT-N66U I was pleasantly surprised to see that throughput increased substantially - double actually (my theory; longer wavelength has better distance, obstacle penetrating and attenuation characteristics).

http://www.smallnetbuilder.com/forums/showpost.php?p=101034&postcount=51

 

[htismaqe]

If you haven't tried the lower channels on the 5GHz band, you might want to.

When I tested the lower channels on my RT-N66U I was pleasantly surprised to see that throughput increased substantially - double actually (my theory; longer wavelength has better distance, obstacle penetrating and attenuation characteristics).

http://www.smallnetbuilder.com/forums/showpost.php?p=101034&postcount=51

I will try it out, thanks.