Optimizing online gaming performance with multiple-cloud and hybrid-cloud strategies — Part two

If you are just joining us, this is the second half of a two-part series from our guest blogger Robert Malnati. Robert is Vice President of Marketing at Cedexis, an Internet performance and monitoring solution provider.

Cloud Strategies for Optimum Gaming

Game performance is affected by many factors, further complicating infrastructure decisions. Building on the free benchmarking of clouds that Cedexis provides, (along with a paid service  we also provide called OpenMix) allows gaming companies and other enterprises to begin to assess where and how they can deliver the best “end-user experience” for dollars spent.

Internap has long realized that relying on legacy protocols or backward looking routing latency assessments to achieve optimal network performance was a wild goose chase. Similar to the way Internap’s proprietary routing algorithm, MIRO, gathers massive amounts of data to improve the performance of its IP, at Cedexis, we have learned that effectively using data can yield unexpected performance gains across multiple locations and clouds..

Two charts of interest here:

Chart 1 above shows how real time data allows for the avoidance of the numerous availability gaps and performance lags that occur throughout a day on any cloud or within any data center’s Internet connectivity. By combining two or more clouds/data centers, your gaming customers can “ride the low latency curve” of optimized performance.

Chart 2 depicts actual comparisons from a Radar Audit, an analysis we do for interested customers using actual Radar data, which shows expected performance gains from cloud vendors they are interested in doing business with. Importantly, note that the different colors display the relative performance of different traffic routing methodologies: Round Robin, Historical Latency data and Real Time Latency data.

Round robin routing is fine for improving availability by maintaining active-active instances, but does little to improve performance. Latency routing using historical data results in limited benefit, as is seen by users of the daily latency data used by AWS-Route 53. Significant gains are found when using real-time data, however, providing meaningful latency reduction for your online game participants, whether they are single or multiple person games (and even for initial downloads).

Optimizing Multi-player Game VM Homing – Multiple player online games provide a more powerful demonstration of the value of a SaaS based GSLB solution like OpenMix. Selecting the optimum location to spin up a multi-player game virtual machine is a great example of the power of Cedexis GSLB (our OpenMix service).

Consider the following use case:

• Three gamers joining their weekly multi-player online game, from Dallas, London and Berlin (blue Icons).
• The game is available from four locations: Chicago Data Center, VA Colocation, New Jersey Cloud and a London Cloud (red icons).
• Optimum location is New Jersey; cloud based on real time latency measures for each location and considering the ISP network serving each player (green icon).

Hence the importance of (1) effectively gathering and assessing real time latency data and (2) choosing among cloud providers, like Internap’s AgileCLOUD, that have appropriately dispersed cloud nodes and system performance.

The gaming business remains robust, but competitive. Having your game downloads delayed or time out due to suboptimal routing or using the wrong combination of cloud node locations/vendors can be expensive. It is my hope that this note served to provide you with the means to be a better decision maker about cloud hosting solutions as you evaluate them for your infrastructure.