BETTER WAN GATEWAY MONITOR - JUST NEED HELP - FREE/DONATION
-
This post is deleted! -
This post is deleted! -
@webdawg I suggest that you begin by writing down the specific logic that you want to apply, and get some general agreement on the proposed logic before you start coding.
The core reason that prior suggestions of enhanced monitoring have failed is because no one has articulated how they would integrate various metrics from multiple interfaces in a way that makes sense.
-
This post is deleted! -
@webdawg The presentation is very high level. You're going to need a much more concrete proposal to reach agreement on. Actionable logic. The presentation is not close to that.
There are a couple of things that I saw in the presentation that I would call out:
-
The assertion that cell based IP is "different" and that you shouldn't use ICMP there. I believe that this is an antiquated view at best. ICMP is a simple fact of life with IPv6, and that makes it a fact of life for cellular IP links as well.
-
The use of the phrase "Even just simple pull < 10K." Bandwidth costs for monitoring are important. Even if done only once per minute, a 10K pull represents a huge increase increase in bandwidth costs for monitoring. I don't think such a cost increase would be welcomed by anyone. There are many people who still pay for service by the MB.
Overall, I would say that the document is focused on sporadic service testing rather than connection monitoring. More appropriate to an NMS than a firewall.
-
-
This post is deleted! -
@webdawg said in BETTER WAN GATEWAY MONITOR - JUST NEED HELP - FREE/DONATION:
I am not looking for someone to agree, I am looking for a dev to tell me if it is possible to implement in a plugin, and what skillsets (lanugages) are needed, and the best way.
The "it" has not been defined. Not having defined the what, it is not possible to define the how.
I'm not trying to be offensive, but this really reminds me of one of my favorite cartoons. A manager is standing in front of rows of programmers at their desks. The caption says "You all start coding, and I'll go upstairs and find out what they want."
-
This post is deleted! -
@webdawg said in BETTER WAN GATEWAY MONITOR - JUST NEED HELP - FREE/DONATION:
THE IDEA
While what "THE IDEA" might be quite clear to you, it isn't clear to others.
The existing ICMP echo request/reply to a specified host is being done for specific purpose. To determine if a link is functional. If the link becomes non functional, certain actions can be taken like declaring the link down, perhaps failing over to another link, etc.. Similarly, if the link recovers from a non functional state, certain actions can be taken like restarting services.
The paper asserts that there is something wrong with the current approach. Honestly, most of the complaints are pretty vaporous. Two hops is bad, I'm interested in three hops. ISPs don't like ICMP. A Verizon engineer said that ICMP on a cellular link is stupid. Really?
Even if we assume that the current approach is horribly broken... it just doesn't work for most of the world, and you want to replace it... with something much better. That much is fairly clear, but after that everything is very, very fuzzy. What are you proposing to replace it with? A 10K iperf test? Who will host the server you test against? How often will you run this test? What action will you take if the 10K test fails? What action will you take if it subsequently succeeds? How does this approach compare with what is currently being done? What are the advantages and disadvantages?
You're not going to get anyone interested until you can answer these kinds of basic questions, and explain why the proposal will be better than the current approach.
-
This post is deleted! -
@webdawg said in BETTER WAN GATEWAY MONITOR - JUST NEED HELP - FREE/DONATION:
@dennypage I don't know then.
I created this so a dev can show up, and tell me if it is possible, and the best way to interface a plugin/what languages I would need.LOL.
Okay. I'll just answer your questions then. You can use whatever language you want to implement your idea. PHP, C, Python are all popular. I recommend that you do all prototyping and development outside the pfSense environment until your work is completely functional and you are ready to integrate. pfSense is currently implemented in PHP, so you will also need to know PHP when you are ready to integrate your work.
-
This post is deleted! -
@webdawg said in BETTER WAN GATEWAY MONITOR - JUST NEED HELP - FREE/DONATION:
I wish I could ban people from threads.
You want to ban the only dev who was willing to take up the conversation with you? Sigh.
I consider myself banned.
-
This post is deleted!
