On 5/10/2019 2:45 AM, Calvin J wrote:
I don't think that it makes sense to make this about me, or you, or
our own companies. I believe that it is better to focus on the thread
topic.
My first post in this thread focused on nothing but the topic, your
first post in this thread focused on me.
I responded to the substance of the comments you made and what the
thread-starter said. I did not involve you personally. That is the way
that civil discussions should work.
I mean, what they're doing is the definition of anycasting. I'm not
sure what industry standard you think disagrees with it. They have
setup multiple vultr vms/servers running bgp and have them routed to a
specific endpoints depending on which IP the source requests.
(https://bgp.he.net/net/92.119.148.0/24#_dns)
Anycast is a network addressing and routing methodology in which a
single destination address has multiple routing paths to two or more
endpoint destinations. Routers will select the desired path on the
basis of number of hops, distance, lowest cost, latency measurements
or based on the least congested route.
If you read that summary and the article, and you look at how the term
is being used here, you'll get a better sense of what I mean. Real game
server traffic has to be forwarded to a central point, and that's not
the same as anycasting, even though the term is being used to mean that.
This was an aside, because propagating the imprecise use of this term
encourages conflating the two. It's not directly relevant to the matter
of whether faking server responses is a reasonable activity and whether
Valve can or should put a stop to it.
You're creating this weird either/or situation. They could be using
vultr's anycast for multiple reasons. Eg. DDoS Mitigation *and*
Routing their own IP block for reliability/ease of movement between
providers *and *because it gives them an added benefit with the server
browser latency.
You are speaking broadly about why someone might use their own IP space
and advertise a prefix at multiple PoPs. Doing that sort of thing is
fairly normal -- it can be called running a distributed network, or
operating a backbone -- and it has its pros and cons (for game servers
on a small scale, mostly cons, but that's another discussion).
Distributed, proxied responses to queries are being used here on top of
the routing behavior (and are dependent upon it as a prerequisite). That
technique is what is unusual. The reasons that have been presented for
these fake responses:
- For DDoS mitigation. This is being done by the server operators. This
is the justification that I refuted.
- In order to deceive clients by making the server look as though it is
lower-latency than it really is or is online when it isn't. This is the
OP's area of concern.
For what it's worth I covered this in short in my original response.
The only time Valve cares about issues like this is when they effect
their profits. This has been shown time and time again, CS1.6 has had
fake player servers for a decade, CS:S had servers that redirected you
when you join, had fake players, TF2 has fake players/idle bots. But
the moment CSGO got skin addons they nuked it, then people came up
with a workaround by swapping tokens and they couldn't be asked to
care about it. Do you think they're going to tackle exploitative use
of bgp?
They /could /do something, but I don't expect much.
The main argument that I currently see for Valve taking action is that
this activity degrades the client experience. But, Valve has chosen not
to continue campaigns against other client-experience-degrading
activities such as the ones you listed. They have generally taken
initial steps to stop them (releasing updates that break the ability to
perform the activity, for instance) but have limited follow-up.
-John
_______________________________________________
To unsubscribe, edit your list preferences, or view the list archives,
please visit:
https://list.valvesoftware.com/