Our Asian Pacific press tour and launch of Omni 2.0 ended this week in Australia where I’m currently typing this from my room overlooking the famous Sydney Harbour bridge. One thing I quickly learned is that you cannot take broadband Internet access for granted. Most, but not all, major hotel chains in Australia have broadband access - at a price. Performance is spotty.
Luckily, Starbucks has a wireless hotspot agreement with Telstra, the Australian Telco giant. I connected from the Starbucks at Darling Harbour and got very good service. Melbourne however, was a different story.
In Melbourne I wanted to work from my hotel room which didn’t have a broadband connection. After dinner I saw this sign in the window of a Chinese shop in the Chinatown district of Melbourne for 10-days of unlimited Internet dial-up for five bucks. This is what is cost for the first five minutes at my hotel’s “business center.” It sounded too good to be true, but what did I have to lose? Besides, after having great success from Tokyo in communicating with our hot-off-the press OmniPeek 2.0 to a PeekDNX engine back at my office in Minnesota, just for fun I wanted to give it a try from a dial-up connection “down under”.
Back in my room, the modem on my Dell laptop worked great with the Australian phone system and I actually got a connection at 53.k Kbps – considerably slower than broadband, but not bad for dial-up. I connected to our VPN and reconnected to PeekDNX back in MN to a session I had started in Tokyo. I was pleasantly pleased to be able to completely control PeekDNX with very good response for remote analysis, troubleshooting, and control of the engine. There were no problems defining/starting/stopping/deleting captures, etc.
Connecting and starting a new capture only took a tad longer than broadband. Scrolling through about a dozen or so packets at a time took only a few seconds and clicking on any one of those packets instantly showed the packet decode and content. Expanding the flow view tree in the Expert in real time also only took a few seconds and current LED activity and problem status indicators worked great. PeerMap operations were instant, as it uses stats periodically compressed and sent from the engine.
All-in-all, not bad for running our distributed fault analysys solution over dial-up, from literally half way around the world. We’re talking about pings showing round trip delay in the neighborhood of 200 ms. On top of that, the data between the console and engine was doubly encrypted - first by Omni, then the VPN. Talk about great security. Oh, and yeah, after the Melbourne dial-up experience, it worked great from the Starbucks Darling Harbour wireless hotspot too.
Hope you all had a wonderful Thanksgiving.