Our good friends down the road got DSL a few years ago so we installed it for them and recently they decided to upgrade to the 2Wire HomePortal 1000HW, an asymmetric DSL modem with built-in 802.11b support.So given that the ADSL circuit had already been set up (the account had been enabled with a PPP over Ethernet name and password defined, and the service had been running) how could this not be easy? One word: SBC.According to the outsourced tech named Bob (whose lousy training, unctuousness and lack of understanding of his job will probably wind up in a future BackSpin column), the problem was caused by a malfunctioning registration server. Three days later, according to other outsourced techs with nice, familiar names like Mary and Joe (but who all sounded distinctly foreign) there was still a problem with the authentication server. Now we got suspicious.The next time we called we demanded second-level support (who are real support people and not, as far as we know, outsourced) and spoke to Eric (not his real name), who was delightfully frank with us about the outsourced techs – “they suck” was his considered judgment. He said the problem had nothing to do with the faulty registration server but with the authentication server, and after resetting the password everything worked. It was a near miracle. We suspect somewhere is the labyrinthine SBC Yahoo DSL installation, something to do with the PPPoE account setup got reset and stopped us from logging on.So with the DSL service up and running and one PC connected by USB, we tried to connect via wireless a Macintosh PowerBook G4 running OS X 10.3. This was another complex and tedious exploration into the nether regions of the deranged ideas vendors have about networks. The first problem was not being able to enter what Apple calls the “password” – what everyone else calls a “WEP key” – because OS X kept telling us we couldn’t get authorized.The default WEP key is printed on a label on the underside of the 2Wire device – it is just a sequence of digits.It turns out that to enter the WEP key in the OS X 10.3 configuration utility you need to enter the key on the label preceded by a “$” so that OS X knows it is a hexadecimal value.Had SBC included the 2Wire manual in the box the company shipped to my friends or even included the instructions in the hideous installation software, life would have been simpler. But no, we had to waste 20 minutes assuming we were doing something wrong then realize we weren’t and had to go searching for an answer online.So now we had a working connection, but wait: No IP connectivity. We checked to see if the G4 was using DHCP and made it renew its lease, but rather than picking up an assigned address from our access point the G4 kept producing some odd address in a totally different Class B network.We tried requesting a new lease multiple times and discovered an interesting effect: If you renew the DHCP lease several times in quick succession the G4 often will respond with a warning it has detected “ping flooding” with the DHCP server as the culprit. Ping flooding, you will recall, is a denial-of-service attack that involves sending a large number of Internet Control Message Protocol echo requests to a target machine. The intention is to keep the target so busy dealing with low-level connection requests that its ability to communicate is partially or completely inhibited.We suspect this problem could be caused by the G4 observing a too-short timeout when waiting for the DHCP server to respond. When the DHCP server does respond, the G4 is no longer listening for DHCP responses and assumes that all the incoming DHCP lease data is a ping flood.Anyway, we couldn’t assign a static IP address to the G4 because apparently the 2Wire doesn’t support such a thing over wireless. After futzing around for another 40 minutes and getting essentially no further, we started another online search to see if the truth is out there.Curiously, there seems to be a significant amount of commentary about the DHCP problem with OS X, but people often appear to get sidetracked by assuming it is caused by factors other than simply not picking up the assigned DHCP lease. There was a happy ending to this saga, albeit one involving an ugly solution. We restarted the G4 and it just picked the DHCP lease. Perhaps the start-up DHCP response timeout is longer than the timeout used by utilities.If you have any clues about this, please let us know at gearhead@gibbs.com. Related content news analysis Cisco joins $10M funding round for Aviz Networks' enterprise SONiC drive Investment news follows a partnership between the vendors aimed at delivering an enterprise-grade SONiC offering for customers interested in the open-source network operating system. By Michael Cooney Dec 01, 2023 3 mins Network Management Software Industry Networking news Cisco CCNA and AWS cloud networking rank among highest paying IT certifications Cloud expertise and security know-how remain critical in building today’s networks, and these skills pay top dollar, according to Skillsoft’s annual ranking of the most valuable IT certifications. Demand for talent continues to outweigh s By Denise Dubie Nov 30, 2023 7 mins Certifications Network Security Networking news Mainframe modernization gets a boost from Kyndryl, AWS collaboration Kyndryl and AWS have expanded their partnership to help enterprise customers simplify and accelerate their mainframe modernization initiatives. By Michael Cooney Nov 30, 2023 4 mins Mainframes Cloud Computing Data Center news AWS and Nvidia partner on Project Ceiba, a GPU-powered AI supercomputer The companies are extending their AI partnership, and one key initiative is a supercomputer that will be integrated with AWS services and used by Nvidia’s own R&D teams. By Andy Patrizio Nov 30, 2023 3 mins CPUs and Processors Generative AI Supercomputers Podcasts Videos Resources Events NEWSLETTERS Newsletter Promo Module Test Description for newsletter promo module. Please enter a valid email address Subscribe