When your router/modem is "here" and you have one or more computers either upstairs or downstairs from that location -- or both! -- life begins to get complex. Hardwiring your network is fast and efficient, but it's often not a practical answer, especially for homes and small offices. Renters may have difficulty convincing their landlords to let them rewire a home or apartment that they don't own themselves. Even wiring your own place may not be fiscally feasible.
Most people who can't or won't hardwire for broadband have an obvious alternative: Wi-Fi. Unfortunately, there can be architectural anomalies between floors or even between rooms that can interfere with Wi-Fi signals, resulting in spotty, or even dead, signals. So what do you do?
Well, you can try using a power-line device.
Power-line networking uses existing electrical wiring to connect your computers and other devices to your network router. The technology behind it allows for multiple signals (and therefore several networked devices) to share the same wires that already exist in your walls without colliding with one another while they're communicating.
Unfortunately, power-line networking has never really caught on as well as it should. It began life at 14Mbit/sec. in 2001 and then bumped up to 100Mbit/sec. in 2005 when hardwire and Wi-Fi were still at 10/100Mbit/sec. and 57Mbit/sec., respectively. Power-line reached a claimed 200Mbit/sec. in 2007, which should be more than enough for our data and streaming-media needs.
There are still a few hitches. Here in the U.S., we tend to run split-phase wiring. The electrical service enters our homes as 240 volts made up of two 120V lines (or legs). Our 120V outlets are derived from tapping off one or the other of those 120V legs. As a result, you may not be able to network devices that are plugged into outlets on different legs. In addition, older wiring and long wire runs can slow down power-line transmission speeds.
We decided to test five power-line devices from Belkin, D-Link, Linksys, Netgear and Zyxel to see both how they compared with more conventional wired and wireless setups and how they compared with one another. (See "How we tested" for details.)
The results? It's obvious that power-line devices rank third in broadband performance. But for those who can't hardwire their homes or for whom wireless networking is less than useful, power-line could be a saving grace.
Which power-line device should you use? Check out the following reviews for some help in making your choice.
Belkin International Inc.'s starter kit provides an access point with one RJ45 connector for your router or modem and a bridge with three RJ45 connectors to which you can attach as many as three additional network devices.
The units are not only smart looking, but they've also been cleverly designed to use either a flush-mount plug (that makes them similar to socket-hogging power bricks) or an included slide-in adapter that provides a couple of feet of power cord for tighter fits. The kit also comes with two Cat5 cables.
There's a button on the front of each device that is the equivalent of WPS (Wi-Fi Protected Setup) security buttons. Pressing one then the other in sequence synchronizes the two devices with 128-bit AES encryption.
While there is a Quick Connect guide and an electronic version of the manual on disc, you probably won't need either. The devices are really plug and play. No drivers are needed. The access point has three status lights indicating power, link status and cable connect, while the bridge has an extra two for the pair of additional RJ45s.
With everything working the way it should, at least electronically, the "room-to-room" clean file transfer completed in 58.82 minutes. When plugged into the same outlet, Belkin hustled the files and folders across the network in a shorter 28.95 minutes.
My 90-foot run needed a modest 33.58 minutes to complete. That time ramped up by about three minutes as each video stream was added to the workload until, at five streams, the file transfer needed 48.51 minutes to complete. (All of the video streams were perfect at all times.)
Except for testing within the same outlet, Belkin was next to last among the five products tested. In this pack, Belkin probably qualifies as an "also-ran," ahead of D-Link Corp.'s device but not an obvious choice.
At a Glance
Belkin International Inc.
Pros: Extremely simple setup, great looks.
Cons: Sparse documentation.
D-Link's approach to power-line networking is somewhat basic. Its DHP-301 starter kit consists of two small white bricks that plug into wall outlets, each with a single RJ45 connector. There is no difference made between which end to attach to the remote device you want networked and which should be the access point connected to your router. Drivers are included on the disc, which also contains an electronic version of the manual.
The prongs on the DHP-301 devices are not polarized. There's no specific way to plug them into an outlet, so you can connect them with the device sticking out either above or below the outlet, leaving the second outlet available for other things. The situation is a bit clunky, but it works.
The room-to-room test was a total bust. By the time the D-Link modules hit the 61-minute mark without successfully completing a clean copy and paste, Windows estimated that it would take another 20-plus hours to complete. I made three attempts in total, but they all produced approximately similar timing. Whatever the problem was that affected the DHP-301, it was significantly more egregious for D-Link than for any of the other power-line kits. After spending 41 minutes on hold with D-Link's tech support and leaving a message when I couldn't wait any longer (I never received a call back), I gave up.
Plugged into the same outlet, the time for a clean copy dropped from nearly infinite to a mere 30.96 minutes. On the more standard 90-foot electrical run, the clean copy operation rose to 49.42 minutes.
Each video stream added to the 90-foot test scenario increased that time by about 2.47 minutes on average, and while all video streams played perfectly, the overall time (61.8 minutes) was the slowest of the five as well.
I'd have a difficult time choosing D-Link over the others, not so much for the speed but because of its lack of meaningful tech support.
At a Glance
Pros: Extremely simple setup.
Cons: Couldn't handle a problematic electrical run, tech support didn't respond.
Cisco-Linksys LLC's kit gets the nod for looks among these five devices. The PLK-300 shares the same piano black styling with blue status LEDs as Belkin's devices but is just a little bit sexier.
The modules use detachable power cords to connect to outlets; one of them has four RJ45 connectors so you can attach multiple devices. As with the other devices, there's no installation to speak of, although there is a Quick Start guide and an electronic version of the user manual on disc.
In most testing cases, the PLK-300 ended as the bridesmaid rather than the bride. It took 50.86 minutes to complete the room-to-room clean transfer; only the Zyxel Communications device, at 50.16 minutes, was quicker.
Connected to the same outlet, Linksys' time to complete a clean transfer plummeted to a much better 29.27 minutes, giving it a second place finish between a faster Belkin and barely slower Netgear Inc. device. The 90-foot test also saw Linksys in second place at 31.94 minutes, this time slightly trailing Zyxel and just barely ahead of Netgear.
Unfortunately, each time a video stream was added to the 90-foot file-transfer scenario, the PLK-300 added 3.28 minutes on average to the completion time. At five streams, the transfer time had risen to 48.35 minutes -- faster than Belkin and D-Link but trailing Zyxel and Netgear significantly. At no time did any of the videos themselves show evidence of stuttering or hiccups, but these results don't bode well for networks that are under a heavy load of mixed media and data.
For lightweight network loads, Linksys is the budget choice here, beating out Zyxel in this category solely on a dollar basis.
At a Glance
Cisco Systems Inc.
Pros: Great looks.
Cons: Highest average increase under load.
The gray and off-gray nodules that make up Netgear's Powerline AV Ethernet Adapter Kit will never grace a display case in the Museum of Modern Art, but who really cares? As with all the products here, installation was basically plug and play. There's only a single RJ45 in each module, and you can only connect one network device to each. The plugs aren't polarized, and you can stand them up or hang them down in your AC outlet so they don't block the remaining outlet at the same time.
There are security buttons on each module that, when pressed sequentially, will set up 128-bit AES encryption between the two devices in case someone is trying to tap into your network. And the three LEDs keep you informed about whether the devices are plugged in, connected to each other and communicating.
As far as performance was concerned, Netgear didn't distinguish itself. The device scored a 56.04-minute completion time for the room-to-room test, putting it in the middle of the pack, behind Zyxel and Linksys but ahead of Belkin and D-Link. the Netgear was also third in the same outlet testing, although -- except for Zyxel -- only tenths of minutes separated the pack in general.
It was the same with the 90-foot testing except when adding video streams. Each stream added 0.98 minutes to the timing on average -- almost half the time it took Netgear's closest competitor, Zyxel. When the dust settled, Netgear actually sported the lowest overall time, 37.76 minutes, when transferring the test packet and streaming five videos.
Performing in the middle actually gave Netgear an advantage: With everyone else taking turns darting above and below its performance level across the various tests, the XAVB101 ended up being the most reliable of the five. Being among the least expensive solutions didn't hurt either.
At a Glance
Pros: Relatively inexpensive, probably the most reliable.
Cons: In performance tests, ran in the middle of the pack.
Zyxel Powerline Ethernet Adapter (PLA401 v2)</a>]
Zyxel's two-part solution to power-line networking includes its Ethernet Multiplug Center (PLA491) and an Ethernet Adapter (PLA401 v2). While the adapter end of the business is pretty much the standard power-brick design, the Multiplug Center has quite a bit going for it: four RJ45 connectors, four 3-prong AC outlets that provide power conditioning (reducing power spikes caused by atmospheric static or even lightning that can travel down the power lines and possibly damage sensitive electrical equipment), two rocker switches to control the On/Off function for the AC lines and the LAN lines, and a security button so you can synchronize the security parameters (128-bit AES) between the two devices.
Except for the security buttons, there's no setup. A trio of green lights on the Adapter module tells you when the unit is plugged in, connected to another device, and communicating on your network. There are three additional LEDs on the Multiplug Center that indicate when devices are plugged into the extra RJ45 ports.
Testing the Zyxel's modules provided a mixed bag of results. Connected to the same outlet, where everyone else zipped along, the PLA491 and PLA401 dragged their respective feet to the tune of 37.2 minutes. That was just a bit more than 6 minutes slower than the rest on average. On the room-to-room test, where almost everyone else lagged, Zyxel sprinted ahead, finishing in 50.16 minutes, a half minute faster than Linksys and more than 8 minutes faster than Belkin.
Zyxel also completed the 90-foot clean file transfer at the head of the pack -- at 30.96 minutes, it was nearly a minute faster than its closest competitor, Linksys. When I added the video streaming to the mix, the times increased by 1.8 minutes per stream on average. Unfortunately, that incremental increased pushed the overall time for a transfer during five streaming videos to 39.96 minutes for an overall second-place finish behind Netgear.
In a lightly loaded network environment, Zyxel would be my choice. Keep in mind that you can save a chunk of change if you buy just two power-line adapters and forego the Multiplug Center.
At a Glance
Zyxel Communications Corp.
Price: Multiplug Center: $174.99; adapter: $74.99; Adapter Kit: $149.99
Pros: Built-in line conditioner; best performance on a mixed AC line.
Cons: High price for the setup tested.
If you looked at our performance numbers, it is pretty obvious that the best thing you can do, by the numbers, is to wire your home. It's labor-intensive, but it's fast -- and as soon as this review is done, I'm plugging the cable back in.
Wi-Fi is a shoe-in for second place. Incidentally, I really would recommend that you upgrade to 5-GHz gear if you're streaming media. It moves everything out of the clutter you'll find in the normal 2.5-GHz Wi-Fi band. (For more tips on getting the most out of your Wi-Fi connection, see "Wi-Fi tweaks for speed freaks.")
If you can't shoulder the costs or hassles of hardwiring and have Wi-Fi dead spots lurking about your home, power-line is a natural. Of the five units tested, I'd have to give the nod to Zyxel for its performance under lightweight loads as well as its AC line-conditioning capability. The drawback is that the price of the products I tested is a little steep compared to the rest. Linksys is the obvious alternative. Overall, however, I'm staring down the middle of the road at Netgear both for its low entry price and overall performance under load.
How we tested
My current configuration is wireless and uses a pair of NetgearWNHDE111 5-GHz wireless adapters as an access point (upstairs) and a bridge (downstairs). For testing purposes, I also used a Linksys WRT600n Dual-Band Wireless-N Gigabit Router and a Linksys EG008W Gigabit 8-Port Workgroup Switch.
I decided to test the power-line devices and the 802.11n and wired networks starting with a clean file transfer of 4,661 files and folders (8.05GB) from a computer downstairs to a computer upstairs. Nothing else would be happening on the network while this transfer was in progress. Then, I repeated that file transfer five more times, each time adding a video stream from a MicroNet MaxNAS RAID with SCSI, which I was using as a media server upstairs, to one of five additional PCs downstairs. That indicated how well each method handled file transfers under load and no-load conditions.
As expected, wired was the fastest across all testing. The clean file transfer needed a meager 6.6 minutes to complete, and that time varied by only an infinitesimal 0.05 minutes for all iterations of streaming video.
Wireless was a bit slower and a bit more varied. It took almost 15 minutes to complete the clean file transfer to the upstairs computer. Each time a video stream was added to the network, that number climbed slightly until it finally reached 16 minutes and 50 seconds when five videos were playing simultaneously. On average, each video stream added about a third of a minute to the final time. The streaming quality of the videos remained the same whether it was one or five playing simultaneously.
In testing the power-line devices, I started with a room-to-room test where I plugged one of the modules into a relatively new (five-year-old) electrical outlet downstairs and the other module into one of the house's original (over 40-year-old) electrical outlets upstairs. In theory, this is the typical home scenario -- for that reason, no attempt was made to discover if the two outlets were or weren't on the same phase leg, since it is unlikely that the typical home owner would know.
I also set up the power-line equipment using a 90-foot extension cord into which one of the modules was connected downstairs; that extension cord was then plugged into the same outlet upstairs as the second module. Basically, this represented a connection through a length of electrical wire in which there were no phase leg, aging or wire condition problems. In a new home, or if you had an electrician run two outlets from your breaker box, you would probably find transmission times similar to these. (This was the setup I used when testing the power-line devices with video streams.)
Finally, I plugged the two modules into the same outlet. In theory, with little to no electrical wire between them, this would be the fastest they could communicate with each other under any network load condition, offering performance under what would pass for ideal conditions.
Don't want Power-line? How to wire -- carefully.
If you're dead set on running wires between the floors or walls of your home, I sincerely recommend that you hire a professional to do the work. Minor intervening items like heating, ventilation and air conditioning ducts and electrical wiring are not only expensive to repair or replace, but a drill bit inadvertently striking a live electrical line can damage a building beyond repair. Also, aside from the safety aspect, having a third party perform the work shifts liability from you to them.
I lucked out, sort of. A few years ago, the local cable TV company added a line for a television at a new location in the living room and tore down some ceiling tiles in the basement to route the cable up from below. One of the missing ceiling tiles is directly above my downstairs network switch, which, by luck, is directly below the upstairs router. With the tile gone, I could see the area I needed to drill and verified that there were no obstructions.
Even so, I used a thin and long drill bit to create an unobtrusive pilot hole, covered it with a lighted flashlight so I could see the location of the hole from below through the darkness of the ceiling, and only then, when I'd verified that everything was where it should be, proceeded to drill the full-width hole. (If you're going to do what I did and use a prewired cable, you may have to "route" that hole out a bit with some drill jiggling to get the connector to fit through.)
Once the cable was accessible, that was it. I just connected it to the downstairs switch and the upstairs router. As all of my computers, switches and routers have 1 Gigabit LAN ports, the immediate benefit would naturally be a tenfold improvement over the 100Mbit/sec. speed available from the wireless setup. All right, maybe not a tenfold pickup, but I was curious to see whether whatever any improvement was worth the angst and possible repercussions of drilling a hole through the living room floor.
Bottom line: It was worth it.