TL;DR
This paper explores the competitive dynamics of "Game as a Service" (GaaS) by modeling how network providers choose their game bundles and subscription prices. Using game theory, the authors demonstrate that providers must balance the "attractiveness" of a diverse game library against the infrastructure costs required to guarantee high-quality streaming for those games.
What Problem Does This Paper Address?
As gaming shifts toward cloud-based models (GaaS), the heavy lifting of computation moves from the user's device to the cloud. However, this creates a massive dependency on the Network Provider (NP) to deliver low-latency, high-bandwidth streams.
The authors identify a strategic challenge: NPs are no longer just "dumb pipes"; they are becoming service aggregators that must compete for subscribers. To do this, they must decide which games to include in their "flat-rate" packages and how to price those packages. Because adding a popular game increases both a provider's attractiveness and its operational costs (paying Cloud Providers for capacity), finding the profit-maximizing equilibrium in a competitive market is non-trivial.
Methods
The authors propose a two-stage sequential game solved via backward induction:
- Stage 1 (Offer Composition): Providers choose which games from a total set $\mathcal{L}$ to include in their bundle. This is treated as a finite game with discrete choices.
- Stage 2 (Pricing): Once bundles are set, providers compete on price. The authors use a modified linear demand model where the number of subscribers for provider $i$ depends on its own price, its competitors' prices, and a novel attractiveness factor ($\Delta_{\mathcal{L}_i}$).
Key Innovations: * Attractiveness Factor: Unlike standard models that only look at price, this model accounts for the specific set of games offered. It includes a "monopoly effect" term—if you are the only one offering a specific game, your attractiveness increases more significantly. * Conditional Popularity: The model calculates the capacity an NP needs to buy based on the probability that a user is interested in at least one game in their specific bundle, preventing over-estimation of costs for overlapping interests.
Datasets
The paper does not use empirical datasets. Instead, it utilizes numerical simulations and a specific case study involving two competing NPs and two games to validate the theoretical model. They assign synthetic values to parameters like price sensitivity ($\alpha$), cross-price sensitivity ($\beta$), and game popularity ($\kappa$) to observe how the Nash equilibrium shifts.
Key Insights and Ideas
- The "Bigger is Better" Heuristic: If two providers face the same capacity costs, the dominant strategy is almost always to offer as large a game library as possible. The increase in subscriber attractiveness generally outweighs the marginal increase in capacity costs.
- The Power of Exclusive Content: The model's quadratic form for game attractiveness highlights that even a niche game can be a powerful tool if it's exclusive. It creates a "monopoly" draw that can offset a slightly higher subscription price.
- Asymmetric Costs Drive Specialization: If one NP has a strategic partnership (lower cost) for a specific game, the competitor may be forced to drop that game entirely to avoid a "pricing war" they cannot win, leading to market differentiation.
- Sensitivity to Popularity: Even games with extremely low popularity (e.g., 1000x less popular than a AAA title) are often worth including because they provide a "long tail" of attractiveness with negligible impact on infrastructure costs.
Key Results
- Analytical Equilibrium: The authors derived closed-form expressions for the Nash equilibrium prices ($p^*_1, p^*_2$) and revenues based on any given configuration of game bundles.
- Numerical Case Study: In a scenario where NP1 had a cost advantage for Game 2, the equilibrium resulted in NP1 offering both games while NP2 offered only Game 1. This shows that cost structures, not just user demand, dictate what content actually reaches the consumer.
- Revenue Proportionality: The study confirms that because demand is modeled as a product of price-effect and content-effect, the content selection doesn't change the math of the pricing game, but it shifts the starting point of the competition.
Limitations and Future Work
- Static Cloud Pricing: The model assumes that Cloud Providers (the ones hosting the games) charge a fixed unit price. In reality, these prices are likely negotiated and could change based on the volume of the NP's demand.
- Fixed QoS: The authors assume the NP always meets the required Quality of Service (QoS). They suggest that future work should treat QoS as a strategic variable—where a provider might offer a "budget" stream vs. a "premium" low-latency stream.
- User Modeling: The demand model is aggregated. A more granular approach using individual utility functions could capture more complex user behaviors, such as "hardcore" vs. "casual" gamer segments.
- Alternative Business Models: The current model focuses on flat-rate subscriptions; the authors note that ad-supported models or "pay-per-play" models remain unexplored in this framework.
Summary updated