ccp-hunter
Latest
EVE dev video talks customer relations and server nodes
As part of a refocusing on developing EVE Online, CCP Games has recently taken us on a video tour of its art, features and core technology departments. This week it's the turn of the customer relations and Virtual Worlds departments as CCP Guard finds out what makes EVE tick. In his visit to the den of GMs, he finds that important stuck queue petitions are answered on average in 15 minutes, a response time much improved compared to a few years ago. Guard chats to several people from CCP's Virtual Worlds department, which is responsible for keeping all the EVE server clusters online and improving performance. CCP Hunter discusses what goes on during EVE's daily downtime and reveals that the once hour-long period has been reduced to only nine minutes per day. We also find out about recent hardware upgrades and the infamous "supernode" deployed for large-scale nullsec wars. Skip past the cut to watch the full video in HD.
CCP explains plan to get rid of EVE's daily downtime
The often-repeated joke that EVE Online is a 23/7 game may soon be put to rest. EVE currently has an hour each day, between 11AM and noon GMT, designated for server maintenance. During this time, important tasks that ensure the smooth functioning of the game are carried out, such as respawning asteroids and load-balancing the server. Over the past year or so, CCP Games has been working to reduce the daily downtime by changing many of these essential downtime functions to happen while the server is up without impacting game performance. CCP's efforts have not been in vain, as the past few years have seen the duration of downtime drop from an average of 50-60 minutes to under 20 minutes. In a new devblog, CCP Hunter explains what has been done to reduce downtime and what the future holds for New Eden's daily slumber. The devblog reveals that most of the current downtime is taken up by server shutdown and startup procedures, with only five minutes of actual maintenance processes per day. Once all the processes required for the daily downtime are removed, the plan is to have the server running permanently.
