backtop

Print E-mail del.icio.us 49 comment(s) - last by Mitch101.. on Jan 7 at 4:27 PM
Marvell's latest Wi-Fi chipset puts others to shame for speed and distance

When it comes to making notebook computers better, the list of things consumers want is long and detailed. A few improvements though are universal to many wish lists for notebook computers. Many consumers want faster processors, better battery life and lighter systems.

Another feature that would make many wish lists is faster network connectivity. InformationWeek reports that Marvell, a maker of semiconductors, announced that it has a new and wickedly fast WiFi chipset it will be introducing at CES 2008.

Currently, the fastest 802.11n WiFi chipsets are capable of a data throughput in the area of 300 Mbps. Marvell's new WiFi chipset puts that data speed to shame with a maximum throughput of a whopping 450 Mbps.

The chipset, called TopDog 11n-450, is able to deliver this blistering performance by using three transmitters and three receivers. In addition to the much faster transfer speeds, Marvell also claims that the TopDog 11n-450 chipset provides a wireless range 500% better than 802.11g chipsets and 160% better than other 802.11n chipsets.

Sameer Bidichandani, senior director of technology strategy at Marvell issued a statement saying, “Marvell is first to market with an 802.11n 450-Mbps solution so manufacturers can deliver industry-leading wireless range and performance to consumers who are seeking faster and more powerful WiFi  capabilities.”

According to InformationWeek, Marvell didn’t disclose customers for its new chipset that will be building consumer devices, but Marvell did say the chipset would be ready for mass shipments in Q2 2008.

Comments     Threshold


This article is over a month old, voting and posting comments is disabled
Real-world speeds?
By OddTSi on 1/3/2008 7:25:17 PM , Rating: 2
So since most current 802.11n set ups that claim 300Mbps speeds only get ~70Mbps in the real world, can we safely assume that this will barely eke out a win over the ancient 100M ethernet?




RE: Real-world speeds?
By SunAngel on 1/3/2008 7:47:33 PM , Rating: 2
I've been fooling around with this whole Wireless-N debacle and I've found that you need to manually set your router and wireless adpater settings to get 300mbps. I had a Linksys WRT350N and an Intel 4965AGN card and was only getting 130mbps. For the life of me, I could not figure out why I was only getting 130mbps. So, when I finally got so frustrated and was going to return both the router and the wireless card I started fooling around with the different settings thinking I could do no harm. I manually set the router settings and made a matched to those with the Intel 4965AGN card and low and behold I hit 300mbps. Keep in mind I had been trying to get this to work for nearly 3 days before I got it to work. I was just so worn out from trying so many different things I returned both of them anyways.


RE: Real-world speeds?
By OddTSi on 1/3/2008 11:31:05 PM , Rating: 2
I'm gonna have to raise the BS flag as high as it can possibly go. Not only on the fact that you claim to be "only" getting 130Mbps with default settings when countless reviews and personal experience with friends' setups shows that if you bust into the 80Mbps range you should probably take a screen shot cause people aren't likely to believe you, but also because you claim to have managed to get 300Mbps "after playing around with the settings."

No person in existence (at least not in the real world, fantasy worlds are a different story) has managed to get the rated speed out of a WiFi setup, be it a, b, g, or even the mimo g routers, yet you somehow happened to do it? Unless you were touching a 99.9% pure, 4 gauge, fully shielded silver wire between the router antenna and the wifi card antenna I'm afraid I'm going to have to label you a liar.


RE: Real-world speeds?
By TomZ on 1/3/2008 11:46:17 PM , Rating: 3
I think it is possible to get a 300Mb "connection," but getting near that "throughput" is a different story. The software will show that high of a connection speed, but in my experience, the actual transfer speeds are much lower.


RE: Real-world speeds?
By omnicronx on 1/4/2008 10:23:41 AM , Rating: 2
quote:
I think it is possible to get a 300Mb "connection," but getting near that "throughput" is a different story. The software will show that high of a connection speed, but in my experience, the actual transfer speeds are much lower.

Yep I think they were talking about two different things, you never actually get the throughput that it displays in windows even through ethernet. Otherwise I would be bragging about my 1000mpbs throughput right now ;)


RE: Real-world speeds?
By TomZ on 1/4/2008 11:31:58 AM , Rating: 2
I agree with that (actual transfer rate less than theoretical), but I think with some of the 'n' wireless gear, there's an additional effect.

I tried out a Belkin N1 wireless router and notebook card for a while, and with that I was able to get 300Mb/s connections pretty easily if I was in the same room as the router. The problem was that, whenever I actually started to transfer data, the connection speed would decrease to some figure much less than 100Mb/s.

I worked quite a bit with Belkin's technical support on that issue. We tried different settings, updating firmware and drivers, and they even sent a couple more routers and notebook cards, but we always got the same effect. In the end, all the hardware was returned for credit - complete failure.

I don't have a broad enough experience base to know if that is generally the case with newer pre-n systems, but from that I learned that the so-called "connection speed" is pretty meaningless.


RE: Real-world speeds?
By seraphim1982 on 1/4/2008 10:04:24 AM , Rating: 2
Wireless-N are definitely able to hit those high speeds. Firstly, the speeds are determined by which wireless-N router and adapter you have and secondly, how you implement your wireless network. The Linksys and Belkin use the Broadcom chipset and only have max 2 (usable antenna), whereas the D-link uses the supierior Atheros chipset and have from 2-3 usable antennas. Secondly, when using a backwards compatiblity with B/G does reduce the speed of the N network. Using a 2.4/5Ghz only N network drastically can increase the speed of WLAN connections. I'm not saying 300Mbps is completely possible with current market items, but 130Mbps is more than DEFINITELY possible. I'm using a D-Link Wireless xtreme Network right now and I am reaching those speeds. This Marvell Chipset though can make 300Mpbs definitely possible


RE: Real-world speeds?
By sprockkets on 1/3/2008 11:45:04 PM , Rating: 2
Heck, I'm trying to get the stupid adapter for a customer to register anything higher than 54mbps on the connection speed in vista with Intel's n card, with that Netgear n router.

Seeing how a certain web site tested that Netgear n router and actual throughput numbers barely were faster than g routers (30-50mbps), I'm not sure how you got 300mbps of throughput. But whatever.


RE: Real-world speeds?
By mindless1 on 1/3/2008 10:18:15 PM , Rating: 2
We can assume 100Mb will still barely win because of the better consistency of the connection. Why are you comparing them though? There is no chance for wireless to ever beat it's contemporary cabled alternative. Now we have gigabit ethernet, or optical lines. Radio < light


RE: Real-world speeds?
By TomZ on 1/3/2008 10:45:56 PM , Rating: 3
Many people are using wireless routers just to connect to the Internet, and since their Internet connection is typically less than 10Mb/s, the difference between wired and wireless is not relevant to many people. And the convenience factor for wireless is a major advantage.


RE: Real-world speeds?
By PandaBear on 1/4/2008 4:04:50 AM , Rating: 2
Given the fact that everyone in Taiwan are now selling G card and router for less than $40, someone has to come up with a way to make money by a "theoretical speed increase" of N.

In real world, most people are limited in line of sight and distance to get any significant bandwidth gain out of MIMO system, let alone the 2x or 3x power consumptions. I barely see an improvement on my DSL when I upgrade from B to G, so I couldn't care less what is N gonna give me. Any serious bandwidth usage should be on wired connection anyways.


RE: Real-world speeds?
By Blight AC on 1/4/2008 8:45:33 AM , Rating: 3
quote:
Any serious bandwidth usage should be on wired connection anyways.


Well, that's just it. Some people do not feel we should be limited in that way.

Until the day that Ethernet connections replace phone jacks in every room of the house, I would like to be able to stream HD media from my Windows Home Server in my Computer room to my Xbox 360 in my Living Room without having to run cable. I can at this point with occasional hiccups, Wireless N should be able to provide me that without the hiccups.

Should I just give up because people are stuck in the 1990's mentality that Wired is the only option for reliablility and speed.. nope, sorry, I'm living in 2008.. welcome to the future, it's wireless.


RE: Real-world speeds?
By HrilL on 1/4/2008 11:22:56 AM , Rating: 2
quote:
Should I just give up because people are stuck in the 1990's mentality that Wired is the only option for reliablility and speed.. nope, sorry, I'm living in 2008.. welcome to the future, it's wireless.


While your point does have some truth to it. The fact is a wired connection will always be better and more reliable then wireless. And what makes these wireless networks work? Wired connections... If you need reliable bandwidth wired in the only option. There is 10Gb Ethernet over copper now that will give you 8Gb/s of actual throughput. Wireless has a very long way to catch up to that. Even to 1Gb Ethernet it has a long way to go. And there are a lot more factors that can degrade a wireless network then a wired one. Also with more and more people getting wireless networks that degrades everyones performance so really imo wireless is only good for 2 things convenience and freedom to move around.


RE: Real-world speeds?
By Blight AC on 1/4/2008 11:40:05 AM , Rating: 2
You know, all that wired speed doesn't really matter when HDD's are only able to read/write at around 100mb/s. Wireless doesn't have to compete with Wired speeds, it just has to get to the point where it's no longer the bottleneck in data transfer.

8 GB/s is great as a network backbone, but as far as the typical end user and endpoint bandwidth goes, that completely overkill.


RE: Real-world speeds?
By joex444 on 1/4/2008 4:59:36 PM , Rating: 2
Just to avoid confusion, what you meant is that hard drives are near 100MB/sec, which is 800mb/s. This is 80% of gigabit ethernet. From the efficiency (ie, throughput/connection speed) of wireless (about 50% max), you will need wireless to increase to 1.6Gb/s before we can max out a single hard drive. That is, before the network is no longer the bottleneck -- and that's if this speed was realized tomorrow. By the time we get that far, hard drives will be much more faster (think Colossal Magnetoresistance heads instead of GMR [today]).

Of course, once you get users like me with RAID5 and RAID10 arrays, all bets are off. I can easily get 100MB/s write speed, real-world throughput, and in excess of 170MB/s read. The RAID10 array is about 130MB/s write, real-world.


RE: Real-world speeds?
By Blight AC on 1/7/2008 11:33:09 AM , Rating: 2
Well, okay.. I missed the whole MegaBit/s, MegaByte/s 8:1 deal. Touché.

However, the most common use of a wireless connection is access to the internet, in that use, your wireless connection is not typically going to be your bottleneck (unless maybe you have FiOS).

I can even stream 1080p video from my Windows Home Server wirelessly to my Xbox 360, using 802.11g. The upcoming (eventually) 802.11n is a vast improvement over that, and should be able to handle all multimedia needs with ease, and the technology for it was ready a few years ago.


RE: Real-world speeds?
By HrilL on 1/5/2008 9:31:22 PM , Rating: 2
I wasn't really talking about end users. I'm talking about high performance files servers with large raid arrays. 8Gb/s = 1GB/s and 2 large arrays can use that much bandwidth.

Well for the typical end user 1Gb/s is overkill still.


RE: Real-world speeds?
By PandaBear on 1/6/2008 12:01:25 AM , Rating: 2
Don't get me wrong, I will be happy to go completely wireless if it is stable and reliable. I use wireless everyday as well, and my point is that if once every 20 minutes wireless give me a hiccup, it already ruins my experience. Hence for me any video related application should be wired, and any high bandwidth demand application should also be wired.


RE: Real-world speeds?
By Blight AC on 1/7/2008 10:08:29 AM , Rating: 2
Well, that's understandable. Although, I must be really lucky with wireless as I don't seem to have anything like that. I watched Season 1, 2 and part of 3 of "The Office" this weekend via Netflix's streaming video service and streamed the video from the Internet to my Laptop, connected by wireless, in my Living Room (with the Router approximately 25' away) and the playback went fine. I was also getting the High Quality stream.


RE: Real-world speeds?
By Yawgm0th on 1/4/2008 4:33:18 AM , Rating: 2
Bandwidth and latency are two very different things. Even a 20mbit connection won't lack bandwidth on a good 802.11g network. But by using wireless you're adding a solid 10ms latency at the very least. Add a few more devices generating simultaneous traffic or some interference, and all of a sudden that wireless is a real bottleneck.

Sure, no difference if you're browsing the web, and fairly little difference if you're web browsing, but it kill online gaming. I've never been able to enjoy a shooter on wireless. I just can't do it. I can feel a 20ms increase in ping.


RE: Real-world speeds?
By TomZ on 1/4/2008 9:01:05 AM , Rating: 2
That's not always the case. I just did a ping test to our DSL gateway, comparing the wired 100Mb connection and the wireless connection. The wired averaged less than 1ms, and the wireless averaged 1ms. So there's no 10ms lag being introduced here.

...and please don't tell me you could feel a 2ms roundtrip difference. :o)

For the wireless, I'm using an Intel 4695AGW IIRC along with a Belkin Pre-N router.


RE: Real-world speeds?
By HrilL on 1/4/2008 11:38:49 AM , Rating: 2
heh where I live every channel of G A and B are bing used. I pick up about 18 networks. When it is peek usage time. say 6-10pm I get about 15-30ms to my gateway over wireless. There is also a lot of 2.5Ghz phones here as well. It all comes down to how many devices and how much bandwidth they are using. Connection speed and packet size effect latency as well. while you might get an extra 2ms using the default packet size with the windows ping command. Playing a game that is sending much larger packets and sending them more rapidly could easily make a 10ms difference.


RE: Real-world speeds?
By TomZ on 1/4/2008 12:02:32 PM , Rating: 2
I can understand how interference can affect latency (and bandwidth), but I don't see how larger packets, or those sent more frequently, would cause the latency to increase, unless there is something seriously wrong with the wireless router.


RE: Real-world speeds?
By HrilL on 1/5/2008 9:25:09 PM , Rating: 2
it doesn't have to do with the router. It is the same with all routers. You can test this yourself. using the window ping command and ping X.X.X.X -l(which is buffer size. Larger packet.)

a 64000byte packet takes 14ms at 100Mb Ethernet. It takes longer for the round trip because it takes more time for each packet to be sent. That is why larger packets have a longer round trip time.


RE: Real-world speeds?
By Martin Blank on 1/6/2008 1:22:43 PM , Rating: 2
Sending a 64KB payload requires fragmenting the payload into multiple frames or packets. Standard payloads for both wireless frames and TCP/IP packets are much, much smaller than that.


RE: Real-world speeds?
By mindless1 on 1/5/2008 4:39:36 AM , Rating: 2
Absolutely, the convenience of wireless connection for my notebook is great.

However, it is not necessarily relevant to think only in terms of what one's internet connection is capable of, now more than ever the typical savvy user has a home LAN which makes the most of the LAN storage, and bandwidth. Internet bandwidth is the least of the concerns since practially anything can meet that.


RE: Real-world speeds?
By Hlafordlaes on 1/4/2008 6:32:16 AM , Rating: 2
quote:
Radio < light


All EM radiation travels at the same speed.


RE: Real-world speeds?
By TomZ on 1/4/2008 9:07:58 AM , Rating: 2
That's an oversimplification of the physics involved. For example, signals travel down a wire at an effective speed which is quite a bit slower than the speed of light.

In any case, the OP's point was that wireless bandwidth is generally always less than wired bandwidth, which has been true so far when you compare bitrates of WiFi compared to Ethernet through the years.


By dyeager on 1/4/2008 8:04:06 AM , Rating: 2
Overcrowding is a frequent complaint of many wireless users. This device is likely aggregating channels across multiple bands. Let us hope is actually listens for other wireless activity while transmitting.

I would rather see wireless speed stabilized and increased performance be put into dividing the spectrum into more channels. This should then result in much better real world performance over time.




Compatability
By daftrok on 1/3/08, Rating: -1
RE: Compatability
By exanimas on 1/3/2008 4:21:34 PM , Rating: 2
It's not a new standard, just wireless N at a new speed. It should be compatible with all thing other 802.11n things are.


RE: Compatability
By PandaBear on 1/3/2008 4:31:14 PM , Rating: 2
It is backward compatible (all n products are), but the problem is the n spec is not finalized and you run a risk of buying something that may not be upgradeable to the final n spec.

If it was not because of CSIRO refuse to assure not to file patent lawsuit, n would have been finalized already.


RE: Compatability
By TomZ on 1/3/2008 7:10:02 PM , Rating: 2
More information on the CSIRO issue:

The Commonwealth Scientific and Industrial Research Organisation (CSIRO) is the national government body for scientific research in Australia. ... Recently, the CSIRO has been actively defending its patent for the use of wireless technologies which are a standard for many modern day laptops. A class action has been filed by US corporations Microsoft, Apple and Dell to renege on paying royalties on the wireless patent filed by the CSIRO in 1996.
http://en.wikipedia.org/wiki/CSIRO

More at: http://www.theregister.co.uk/2007/09/21/802_11n_pa...


RE: Compatability
By sprockkets on 1/3/2008 11:30:15 PM , Rating: 2
They've sued and barred Buffalo from selling their wireless
routers here in the US. Too late.


RE: Compatability
By SunAngel on 1/3/2008 4:58:51 PM , Rating: 4
I'm going to go out on a limb here and say Wireless-A has not flopped. At the time Wireless-A was ratified and introduced Wireless-G was well established, thus many (people) had Wireless-G equipment. Wireless-A served a different purpose than Wireless-G. Wireless-A was for the transmission of media content because of it different band usage than Wireless-G. Both transmit at 54mbps, but Wireless-A is on the 5GHz frequency and Wireless-G is on the 2.4GHz frequency. I would further say Wireless-A's ability to stream content is (a) little stabilier because of the less crowdedness on the 5GHz band. However, I found that to be a little misleading. When I got a 5GHz wireless telephone I found that when the telephone rings I get a dropped connection, but it returns within about 20 seconds. Same thing used to happen when I had a 2.4GHz telephone and was using the Wireless-G band. I've a resort back to a 900MHz telephone to keep my internet connection to keep from dropping.


RE: Compatability
By m1ldslide1 on 1/3/2008 5:12:52 PM , Rating: 2
I have to agree that 802.11a isn't the flop that some depict. Just because it doesn't live in notebooks doesn't mean that there aren't other valid uses for it. I know that Cisco has outdoor mesh AP's that use the A radio for backhaul, and I would imagine that other vendors do the same.

What I want to know, is that if this chipset uses 3 separate transmitters and receivers, does that mean it creates 3 different associations with the AP or does it use a GLBP-type MAC spoofing mechanism for a single association?


RE: Compatability
By SunAngel on 1/3/2008 7:34:19 PM , Rating: 2
I would think so. The Linksys WRT600N has two separate radios. Both transmit at the same time, thus you have two separate SSIDs.


RE: Compatability
By PandaBear on 1/4/2008 3:54:32 AM , Rating: 2
I don't think n spec uses multiple channel for that. What I understood is it uses the same channel but due to the timing and spacing between antenna, the overlap will not interfere between tx and rx (spatial multiplexing).

Sort of like antenna T1, T2, T3 all send out different data at the same time, but are 1 frame apart, and antenna R1, R2, R3 will receive T1 first, then T2 mixed with part of T1, and T3 mixed with part of T2 and T1. R2 will have to subtract T1 to get the original T2, and R3 will have to subtract T1 and T2 to get the original T3.


RE: Compatability
By Mitch101 on 1/3/2008 5:16:48 PM , Rating: 2
I fully agree with you. I have no 2.4ghz devices running and no headaches. A few of my geek friends are running WiFi-A because its faster and more reliable than G because of the 2.4ghz overcrowding. WiFi-A is also rarely hacked because no one is looking for it.

I had to purchase a 5.8ghz phone because of all the 2.4ghz overcrowding. Too many times with no dial tone because all the channels are being used. I have heard of some people also going back to 900mhz because of everyone running to 2.4ghz. I'm also convinced that the additional range of 2.4ghz devices will only add to the overcrowding problem.

For me I ran Gig-E through the house because I don't like delays or network issues that most people in my area have with Wifi Bandwidth. Wifi-N is still overpriced. A day with a $39.99 spool of cat 5e and $15.00 worth of covers and jacks, dry wall knife, and some Spackle and paint and its done. No network bandwidth issues ever and the Tivo's, X-Box, Media Extenders are all happy.

On a side note I am glad to see this tech move forward. Maybe one day we can choose the frequency, 900Mhz, 2.4Ghz, or 5Ghz to hopefully eliminate this overcrowding.


RE: Compatability
By mindless1 on 1/3/2008 10:13:29 PM , Rating: 2
I too was quite willing and happy with the results of stringing gigabit ethernet cable, but I wouldn't do without my 11g for the notebook.

I don't find my 2.4GHz cordless phone has any problems due to using the notebook (wirelessly), and would suspect if yours does (unless you live in very close proximity to others with 2.4GHz gear, I mean in same physical building) then it might just be the phone's fault.

Regardless, you ought to look into DECT6 phones. Longer transmission range plus (typically) much better battery life make them more desirable than 5.8Ghz, and the band is even less crowded than 5.8GHz.


RE: Compatability
By Blight AC on 1/4/2008 9:30:27 AM , Rating: 2
Yep, great for you, but what about the millions living in apartments?

That, and as was mentioned by the other post.. Laptops.

I also find it funny that you can't stand wireless issues, but purchase wireless phones.. if you going to tout wired connections, go all the way man.


RE: Compatability
By Mitch101 on 1/7/2008 4:27:08 PM , Rating: 2
I enjoy wireless products but 2.4ghz is overloaded. I can hit 20 Wifi Points from my house alone and would bet at least 40-50 2.4ghz phone setups. Keep in mind some households will have more than 1 because they have teenagers who might have their own phones also.

switching out the 2.4ghz devices I have not had any problems.


RE: Compatability
By Fox5 on 1/3/2008 5:51:36 PM , Rating: 2
I'm pretty sure wireless A was before G. Wireless A was never intended for consumer adoption though, and was focused more on business use. At the time, I think the FCC was trying to keep the 5Ghz band relatively clear, so wireless A never had a chance of reaching the penetration levels of b and g. I believe that N is 5ghz though, so I suppose the FCC must have changed its mind.


RE: Compatability
By TomZ on 1/3/2008 6:58:44 PM , Rating: 3
You're right - 'a' was before 'b':

http://en.wikipedia.org/wiki/IEEE_802.11

I think in reality, the reason that 'g' took over market share compared to 'a' was that it offered the same bandwidth and greater range than 'a' did. This is because the higher-frequency 5GHz 'a' signals are absorbed a little more by walls, doors, and other obstacles compared with the 2.5GHz 'g' signals.


RE: Compatability
By PandaBear on 1/4/2008 3:57:24 AM , Rating: 2
n is 2.4Ghz, with MIMO


RE: Compatability
By TomZ on 1/4/2008 9:03:39 AM , Rating: 2
...and 5GHz. 'n' uses both frequencies.


RE: Compatability
By AstroCreep on 1/3/2008 5:18:37 PM , Rating: 2
Maybe I'm missing something, but how did 802.11a 'Flop'?


RE: Compatability
By mindless1 on 1/5/2008 9:19:29 AM , Rating: 2
One might say it flopped simply because manufacturers making mass goods for the consumer market jumped onto the 11b and g bandwagon so the lower cost (versus A) caused mass adoption. Anybody remember the old 900MHz gear? Same thing but even moreso, and mainly used for bridging instead of clients.


RE: Compatability
By Mithan on 1/3/2008 6:53:54 PM , Rating: 2
Wireless A is more for VOIP type connections and is used.


"Let's face it, we're not changing the world. We're building a product that helps people buy more crap - and watch porn." -- Seagate CEO Bill Watkins

DailyTech Poll
Which web browser do you use on your primary personal machine? 






44 Comments







botimage
Copyright 2009 DailyTech LLC. - RSS Feed | Advertise | About Us | Ethics | FAQ | Terms, Conditions & Privacy Information | Kristopher Kubicki