backdrift

I merely work under the assumption that 343 hosts the servers and equipment. If they're outsourcing that to Azure, that's different. However that (as a customer) should be and is completely transparent to us. Similar to how ISPs sell internet to lower or smaller tier ISPs in remote areas. The customer is not going to complain to TWC for Middle-of-nowhere Town's ISP service. They're going to complain to Middle-of-nowhere. They should have an SLA with their provider the guarantees a certain service delivery, as should we as customers. However the Azure cloud platform is very similar to Cisco's Nexus 9k platform, that does ACI for dense performance. It really comes down to a reliable load balancing perspective. It's hard to blame it on network performance or routing, since (i can go on for days about EGP protocols like eBGP at the backbone), because they are coded to deal with this sort of loss. Besides, BGP peer idle and transitions states work in a matter of milliseconds, and usually result in a 1/1000 packet loss. Halo is equipped to deal with that sort of loss. Ultimately there is something in the core environment underperforming. If 343 leases that space for their services, they need to work with the vendor to stay true to the SLA they likely purchased for the service. I did try to enjoy my time with Halo 5, but unfortunately the lag made it unplayable. Microsoft was kind enough to refund the 100$, after they realized I exhausted all troubleshooting options which were as follows; Open NAT on Xbox Static IP & default route pointing out a dedicated internet circuit (using 8.8.8.8 as DNS). Running and both 2.4ghz and 5ghz wireless. All necessary port trigging and ports forwarded on the router. In a corporate environment and home, the Xbox was placed in a DMZ so no firewall would block. A NAT was configured fully open from 1-65535 TCP/UDP ports. Direct connections to overture in Verizon and TWC, as well as home modems. Cat 5, Cat5e. and Cat6 cabling testing with all above variables. I did however limit the connectivity problems to one circumstance. When ever I was in a party and I was in matchmaking. Only then did I receive rubber-banding, and other sorts of lag-related ailments.

Talk to Microsoft. I opened a chat session with them yesterday evening and they were pretty quick to provide me a refund for the Digital Deluxe copy. Granted it was due to the horrible lag and other issues that made the game unplayable after countless hours of troubleshooting.

From some one who has spent 10 years in the Networking field, allow me to respond to what you said here. 1.) Server usage/traffic is likely due to poor QoS from their internet circuits. Which means they need to re-prioritize what traffic is being identified as EF. Also they are more likely utilizing F5 load balancers, which is they are getting hammered or not able to handle the traffic, that's actually a 343 problem (unlike what you said earlier). They need to implement additional devices, which I am certain they have the funds for. Then they create additional VIPs that will do a round-robin request and allow traffic in to additional servers that are in a farm. If servers are underperforming, 343 needs to fix it. 2.) Redirect communication lines? Are you referring to load balancing? If so, see #1, because that is still a 343 issue. 3.) Despite popular belief, that's not how ISP service works. ISPs use a similar concept to MPLS, which is tunneling either via layer 2 VPN or layer 3, or through dedicated DLCI in frame relay to your modem. The modem acts as a demarc and termination for their service. That's why they manually register your MAC, so that it can static assign a layer 3 address, and tunnel your traffic to the backbone. Your neighbors do not affect this, unless you're sharing internet. 4.) This is true, inhouse devices impact internet. However a constant TCP session that is open after the TCP handshake usually has an ACK that is resends with no data if the session is constant. Which doesn't even equal 1Kb. The header in a TCP ACK is so minimal, it's irrelevant. On a 50Mb circuit, where TCP port 80 and 443 traffic is all that is being used on 1 or 2 devices outside of the Xbox, you can rest assured that it's not the internal network. 5.) Server downtime or redirect will impact server connections, sure. However there are about 10000000 ways to guarantee service on F5 when you have a VIP that redirects to multiple farms. This is also a 343 issue they should be fixing. 6.) If your servers, or F5 load balancers or other type of load balancer, cannot keep up with requests that is yet another 343 issue they should fix. 7.) Firewalls do suck. This is the only valid point I will agree with. Port Forwarding is necessary, even if you don't have a firewall. This is also to recognize that overloaded TCP headers designated to a certain MAC/IP, regardless of what happens in a handshake or retransmit, the traffic knows what port it is destined for. That way the retransmit, during data loss, service delivery can react quicker. 8.) See #7 Also, what is "Network Coding"? Are you talking about Network Configuration? Configuring a switch, router, ASA/Firewall is configuring, not coding. While this might seem like I am picking on you, if you want to proclaim you're a network admin, engineer, or architect, you better expect some one who is also in the same field to be fairly critical of what you say. At the end of the day, we are all having issues. I do not anticipate that it's a TWC issue, I anticipate it's a 343 issue. Whether it's overloaded circuits, bad routing, unresponsive servers, I don't care. I received a refund for my Digital Deluxe purchase. I would highly suggest all of you who cannot successfully resolve their issues, contact Microsoft Support and demand a refund as well. The loss of revenue will be the only way to induce change in a business.