January 31, 2003
Verizon vs. RIAA
I have a couple of thoughts about the raging Verizon/RIAA court battle.
1) I'm not convinced that Verizon's stated position on the latest appeal is the whole story. The Register has been quick to congratulate Verizon on their stance, but I'm a bit more suspicious of Verizon's motives. Publicly-held companies will rarely burn through cash for purely altruistic reasons -- their shareholders won't stand for it. There must be some other reason(s); probably ones that ultimately affect their bottom line.
2) Suppose Verizon has to reveal the identity of the mass downloader. What happens then? How will the RIAA proceed? Probably by leveling charges, serving subpoenas, and getting warrants issued, without being required to prove anything. If it gets to this point, the RIAA (and anyone else who wants to [mis]use the DMCA) has an incredibly potent beatstick against average people. Most ordinary folks can't afford the legal hassles that could be dropped on their head at the whim of some faceless corporation. The mere appearance of possible copyright infringement, without any proof, should not be sufficient to permit the circumvention of due process of law.
[Later...] More discussion over on Copyfight.
January 30, 2003
More "Wi-Fi Zone" thoughts
Just came across Scott Rafer's blog, Provisioning Soweto. There are a couple of articles on his site that tie in to my ongoing thinking about the "Wi-Fi Zone" initiative of the Wi-Fi Alliance.
One article from Rafer's site, titled "Cooking Wi-Fi Zone", raises several complaints about Wi-Fi Zone:
1) Licensing scheme
2) Site survey requirement
3) Backhaul requirements
4) Adoption of paid hotspot networks
I also think that the Wi-Fi Alliance's licensing terms are too steep at this point. A minimum of $100 per year is a lot to pay for small venues. Is it worth that for a quality guarantee? What does T-Mobile say about their wireless hotspots?
The site survey requirement seems to be a checkbox on a form somewhere. No matter what the size of your venue, you need a site survey? At least it's fairly simple to do, and you only have to do it once, unless the Wi-Fi Alliance receives enough complaints -- they can demand another site survey.
Tying backhaul requirements to the number of access points seems strange at first glance. Even second glance. But, APs are much easier to count than users. And a 128 Kbps backhaul speed insures some level of throughput. About the same as dialup, if there are very many users :)
I am also skeptical about the viability of a paid access business model. I agree with the thinking outlined in the second article on Rafer's blog -- "Like charging for salt and pepper".
In that article, a venue owner says he pays $60 a month for broadband access (1.5 Mbps), which he uses anyway, and he gives the rest away.
Giving away bandwidth, especially if a venue already has broadband, is a small cost of doing business, and a big differentiator for customers. And a big plus -- it's a lot less hassle for owners and customers.
January 28, 2003
802.11g interoperability
More information about 802.11g playing nice, from Unstrung. There are 2 interoperability questions: between 802.11g devices from different vendors, and between 11g and 11b devices.
Interoperability between different 802.11g chipsets is certainly an issue. However, they are only typical interop questions, and there will be bumps along the way as they are ironed out, but there are no showstoppers.
Interop between 11g and 11b is another, more complex issue. The root cause of this problem is that 11b devices cannot "hear" an OFDM waveform, which 11g uses for higher speeds.
802.11 uses a mechanism called CSMA/CA, which stands for "Carrier Sense Multiple Access with Collision Avoidance". This is a complicated way to say "listen before you talk". This differs from wired Ethernet, which uses CSMA/CD -- "CD" means "Collision Detection", or "listen while you talk". 802.11 cannot listen while talking, because the receiver is overloaded by the transmitter.
So, when an 11b device tries to "listen", what if an 11g device is already transmitting a message using OFDM? The 11b device thinks the air is clear, and transmits. Both messages are corrupted to some degree, but the AP has already locked on to the OFDM message, and the AP will not be able to ACK the message from the 11b device.
This is particularly problematic when the 11b device has just been enabled, and it is actively scanning across channels looking for an AP. If an 11g device is hogging the AP, the 11b device will have little chance to "get lucky" enough to transmit a Probe Request successfully. Passive scanning may be more successful in locating APs, but the same problem exists when the 11b STA tries to authenticate or associate.
Using RTS/CTS can help the problem, but not eliminate it. The RTS/CTS sequence can be heard by 11b devices, but it is a very short time interval in the RTS/CTS/Frame/ACK exchange. A max-size packet, sent by an 11g device using 6 Mbps OFDM, takes slightly more than 2 milliseconds just to transmit the data. RTS, CTS, and ACK frames are short management frames without any data.
In the presence of high 11g traffic, an 11b device that is changing channels has a high percentage chance of listening in the middle of the OFDM data frame -- which it cannot hear.
RTS/CTS is an optional feature in 802.11. RTS/CTS adds additional overhead and reduces throughput. However, RTS/CTS alleviates the hidden node problem. And RTS/CTS is helpful in networks that mix 11b and 11g clients, but it does not fully address the interoperability issue.
For more information about 802.11g:
January 27, 2003
Wi-Fi misconceptions
Just read an article on ZDNet about Wi-Fi, by way of Glenn Fleishman's Wi-Fi blog. The ZDNet article raises some valid points, but contains so many misconceptions and inaccuracies that, as a whole, the article does more harm than good.
First, the good stuff. As mentioned in the article, 802.11b equipment is much faster than almost all internet access types available for home users. Upgrading to 802.11g for home use will not affect applications that move data in and out of the house (via the internet). Applications that use the higher throughput of 802.11g will be those that move data around the home, not in & out. Similarly, in an office setting, 802.11g will speed up access on the internal corporate network, but probably not for anything that goes through the internet bottleneck.
The article also has a nice, concise summary of WPA (Wi-Fi Protected Access).
Now, the content I have issues with. I'll list them in the same order they appear.
- 802.11g and 11a are 54 Mbps, not 55. (Small, but the devil is in the details, ya know :)
- Multiplayer games are not typically bandwidth hogs. They are designed so those poor folks with dialup can play, to maximize the potential player pool. However, MP game traffic is time-sensitive; you don't want a 300 ms ping time in a first-person shooter. Bandwidth never hurts, though :)
- 802.11g will be backward-compatible with 11b. The main issue is that interoperability between different 802.11g chipsets is not guaranteed for the newer 11g operating speeds and waveforms. The IEEE has not yet formally ratified the additional OFDM PHY modes and necessary MAC extensions for 11g, but the 11b modes will be supported by 11g. The purpose of the Wi-Fi alliance is to insure interoperability, there are 11g activities under way, and 11g will be included in Wi-Fi certification once the IEEE ratifies it.
- 11b and 11g protocols can run at the same time, and one 11b card will not force the entire network to 11 Mbps. 11g devices use RTS/CTS messages (which 11b devices can hear) to reserve bandwidth for OFDM messages (which 11b devices cannot hear).
- There will not be tri-mode cards, just dual mode. 802.11g is a superset of 11b. The new dual mode cards will be 11a/11g, running at 54 Mbps in either 2.4 GHz or 5 GHz.
- 802.11a is more secure because it has shorter range? Please. The article mentioned that this is a weak argument, but why was this even brought up?
January 25, 2003
Slammer running rampant
Just catching up on SQL Slammer. Apparently it's a worm, specific to Microsoft SQL Server 2000, that is bringing some parts of the Internet to it's knees.
And this just 10 days after Microsoft reported progress on Trustworthy Computing after the first year. Ouch.
Introducing Bosco!
This is Bosco, our Sphynx. He took a break from sunning in the window, and came over to say hi. So, I picked him up for a while, and then we played fetch. Yes, he fetches his cat toy.
I'll introduce our other furry companions later.
"Consumers" and DRM
DRM pisses me off. And I think I'm starting to realize why.
In the context of consuming/producing, DRM specifies (and enforces) what type of consumption is permitted.
It is nothing short of unmitigated hubris to impose limits on the legal ways a product is consumed.
Industries looking at DRM as a solution to their problems had better look toward a more permanent solution. As production and distribution costs for digital products plummet, media companies need to focus on their core role: serving as an intermediary. Until they return to truly value-added activities, the ship will continue to sink.
Constraining production methods is a sustainable practice only when a true monopoly is present. Attempting to constrain consumption patterns on goods that are rapidly becoming commodities is suicidal.
Media companies, serving as intermediaries between artists and customers, need to focus on adding value, instead of trying to preserve an outdated business model through technology (DRM) and law (DMCA).
There, I feel better now. Whew.
January 24, 2003
Online and rolling
It appears the blog is up and running.
After I called my ISP to set up webhosting, it took me about half a day to get the blog going.
Congratulations and appreciations to Ben and Mena Trott, for putting out Movable Type. A slick product, indeed.
First entry...
My first plunge into the world of blogging. Why do I feel like I'm on a high platform with rubber bands tied to my ankles...?