Fabio FLO Floris @FF_Flo

@archerace27 @GlitchtechES Congrats 🥂 🇮🇹

Genuinely - what "lie" are you referring to exactly? We are in the process of writing a more technical blog post to share more transparently they way we are thinking and operating within our AC effort - I would love to know more what is the exact concern/challenge you are sharing here

@sball9999 @FACEIT_Darwin @FACEIT How it was ?

You need to design a piece of hardware capable of behaving perfectly as it supposed to behave while doing something else; which is something very complicated to accomplish which is the reason why AC have been detecting these type of cheats since a while. IOMMU is recent and definitely very strong upgrade in terms of protection - however detecting DMA w/o IOMMU is not a "claim" is something technically achievable every time a device based cheat is not perfectly designed (like any other type of cheats)

I genuinely don't know the exact answer i.e. whether we are actively triggering an ID Verification (IDV) based on absolute stats value only and I'm happy to share this internally if you have an account to pass over. What I can say is that the system of triggers to enforce IDV we have today is far from being completed. Since last November, we significantly expanded the amount of information we use for "fingerprinting" players, but we also need to observe patterns for some time to make our data strategy for identifying suspicious accounts stronger while avoiding false positives before pushing live all the enforcement at scale = we need to leave some "gates" open for some time if that makes sense. This is also true for the actual IDV process, which today can still be abused in certain instances (i.e., verified accounts being bought online). On both topics, I think we are not far from releasing a first set of updates.

@HarryStebbings .@RayDalio thoughts ?

@fREQUENCYCS @ZPostFacto @willste92 @CounterStrike @FACEIT @FACEIT_Darwin Appreciate the tag ; as @ZPostFacto mentioned - it’s very possible the issue is happening in the route between client and server. Happy to look into @willste92 logs to confirm we may be able to reach out to the ISP or routing partners

@willste92 @fREQUENCYCS @CounterStrike @ZPostFacto @FACEIT hey can you link me your FACEIT profile?

@abp_cs sir! can you open DM ? - from FACEIT

fair point but since the correlation we have between matches being streamed VS ddoss attacks - we tend to assume is someone having "fun" in attacking and watching players experiencing issues; but you are obviously right and it was misleading for me saying the attacker definetely snipped the IP from the streamer view compared to other scenarios mb

@TheWetBreade @FraqKTR @FACEIT @FACEITcs luckily its not the case but there is definetely an higher ratio compared vs matches played (compared to EU for example) and whenever a "match" is attacked unfortunately its likely you have another 50-80 players impacted while playing on the same machine if that makes sense

Hey, no worries at all — totally fair question. Datacenters have DDoS protection which are tuned to detect some type of data attack (type of packet and size etc etc) in this case the attack manages to stay under radar from those thresholds/detections while still being effective enough to degrade the experience. The tricky part is when the server IP is exposed the attackers can bypass protections by targeting the match directly with low-scale specific attacks like in this case. It’s 100% still super frustrating from an UX standpoint and we should work with the datacenter to add this detection. Regardless the most effective thing would be stop exposing the IP in full. Hope it makes more sense

I have reviewed some available public posts and complaints about the freezing and lag issues you described, where all players are impacted simultaneously. These are all DDOSS where sometimes the incoming traffic triggers a DDoS alert. In other instances, as shown in the image, a DDoS attack is significant enough to disrupt gameplay but not large enough to trigger an alert. Since many public clips of these issues come from streamed matches, make these matches/machines more likely to be targeted by malicious traffic. To address this, i guess we should consider moving some competitions to a client-based user flow that masks IP addresses completely from the UI

@Robosquad_cs @wildlifedudley @FACEIT What location you personally have been experiencing those crashes? And what frequency are we looking at ?

From another reply because is late night 🥹 here "I fully understand that you may be referring to issues beyond the general "bad feeling" players are experiencing (compared to the past and official servers). Regardless, my point is that recurring issues affecting the machine (CPU, memory, etc.) are easy to detect, and updating, upgrading, or replacing the machine or vendor remains a simple decision, including testing a different data center location to improve the overall experience. However, addressing the broader issues reported in surveys and post-match reports isn’t as straightforward as machine-level fixes; we need innovative solutions for network-related problems. Machine crashes or DDoS attacks can still occur, but these are isolated incidents that have always existed, even in the past.

I fully understand that you may be referring to issues beyond the general "bad feeling" players are experiencing (compared to the past and official servers). Regardless, my point is that recurring issues affecting the machine (CPU, memory, etc.) are easy to detect, and updating, upgrading, or replacing the machine or vendor remains a simple decision, including testing a different data center location to improve the overall experience. However, addressing the broader issues reported in surveys and post-match reports isn’t as straightforward as machine-level fixes; we need innovative solutions for network-related problems. Machine crashes or DDoS attacks can still occur, but these are isolated incidents that have always existed, even in the past.

If you see all players spiking latency etc etc I agree it’s more upstream the issue but again is always more likely a network issue (a big congestion - ddos etc etc) what I am referring to here is a wider perception many users have been sharing on the servers not feeling as smooth as they were before CS2 - in particular compared to the official servers that being hosted on the steam datagram relay make any network related issue almost irrelevant. Bottom line - it would very simple for us upgrading a machine, changing a vendors or testing a different server configuration; between FACEIT and ESEA folks we had quite a bit of experience on that front , unfortunately we are dealing with something that cannot be solved so easily at the machine level

Official servers being hosted on steam datagram relays + ping based matchmaking - make this issue almost irrelevant outside of community servers - you can ask around but this is an issue that is recurring in pretty much any project running on community servers. We fully understand the frustration and it would be a very stupid decision for us not simply doing what we have been doing for years: upgrading machines / charging data center / changing configurations these are all things extremely simple to do. Anyways hope you guys will feel the difference once we go live with what we have been working on. It’s not too far we start testing live with players

trust me upgrading/replacing a server machine is extremely fast and cheap :) Players experience the same lag because CS2 is not as optimized as GO in compensating latency discrepancies of players and therefore the only way we can solve this issue is working on the network side of things (which is not as simple as upgrading a server machine). Makes sense ?