Cathy Zhou wrote: > Dan, > > Thank you for your comments. See inline: > >>> * Vanity naming >>> >>> Today, the network names are tied to the underlying network hardware >>> (e.g., bge0, ce0). Because configuring the system always requires >>> network interface names to be stored in a wide range of >>> configuration >>> files, being able to give a meaningful and consistent vanity name to >>> network interfaces will make network administration much easier. >>> More >>> importantly, vanity naming will prove especially useful for machine >>> migration and Dynamic Reconfiguration. >>> >> I wonder if there is a better way to word this paragraph. When I read >> it the first time, I thought "How would giving an interface a >> meaningful vanity help with interface names scattered through a wide >> variety of files? Wouldn't we have to keep track of each instance in >> each file?". After thinking about it, I could see how having a vanity >> name would help (i.e. indirection). It feels like you're missing a >> step in explaining how vanity naming would be helpful. >> > Right, vanity naming does not help to track those configuration files. > But carefully-chose names can help to make the configuration meaningful. > Is that not clear? Which step do you think I missed? >
On my first read through your paragraph I didn't understand how giving an interface a meaningful name helped an administrator in managing the configuration when the configuration is stored across numerous files. Since we're not managing or helping him track those files, is it because it becomes easier to search for what files contain an interface's configuration? Is it easier to implement an indirection layer? Does that help? >>> - By what criteria will you judge its success? >>> >>> The project will be complete once the following requirements have >>> been met: >>> >>> * Must provide a consistent model for network interface >>> administration: >>> >>> - All network interfaces are administrated by the same set of >>> commands. >>> >> >> Even third party interfaces like Syskonnect? >> > Anything special of the Syskconnect interfaces? I would think they are > just legacy devices which can be managed by softmac then administered by > dladm. > OK, I remember we got comments about Syskonnect and other third party NICS during the VLAN observability PSARC review, and I didn't want that to happen again. >>> - Is the project internationalized and localized? >>> >>> N/A >>> >> >> What happens if a user wants to name an interface using characters >> that aren't in 7-bit US-ASCII? Is that even possible? A few examples: >> >> - a Chinese user naming an interface using a Chinese character set >> (like Big5) >> - Spanish users typing names that include tildas or accent marks >> in a UTF-8 environment >> - Russian users wanting to use names with Cyrillic characters? >> > Meem answered this. > I saw, and I wouldn't be surprised if we get pressure to change this (see my response to Meem). Dan
