• Proposed a long-reach optical access network architecture (LCQ-LR-OAN) that utilizes backup resources for providing the protection and alleviating network congestion. The LCQ-LR-OAN can simultaneously support different types of users with different QoS requirements. • Proposed a partial protection strategy that provides protection to all shared optical resources while keeping the unshared optical resources unprotected. We also developed survivability and network congestion alleviating mechanisms using a single set of optical resources. • Considered user priority (e.g., FTTH, FTTC, FTTB and FTTO) with reference to providing the protection. • Proposed a novel DBA scheme (EBA) to allocate the bandwidth of the backup OLT to the LR-PONs during the network congestion periods. • The LCQ-LR-OAN requires lower cost compared to that of the other existing LR-OAN architectures. The LCQ-LR-OAN architecture does not depend on any specific PON standard, viz., EPON, GPON. Thus, it can be used with all the PON standards. In this paper, we explore a low-cost and quality-of-service (QoS)-aware long-reach optical backhaul for the wireless and fiber-to-the-x (FTTx) networks. We exploit the reserved and idle optical resources to provide protection against failures and offer extra bandwidth to the users in the network congestion hours. We use a single set of optical resources to both provide protection and reduce network congestion, thereby incurring the minimum total expenditure. For the practical network scenarios, the simulation outcomes reveal that the proposed architecture can reduce 11 - 79 % average packet delay in the network congestion hours. It offers higher connection availabilities, i.e., five-nine availability ( ∼ 0.99999 ) with full protection and four-nine availability ( ∼ 0.9999 ) with partial protection. Meantime, it requires ∼ 30 % less capital expenditure cost than the other existing benchmark and state-of-the-art long-reach optical backhaul architectures. The proposed architecture is validated by simulation studies and practical data.