Rollup is an emerging off-chain transaction processing solution for blockchains. With the continuous development of applications, the need for interoperability among different types of Rollups is increasingly growing. Existing cross-Rollup interoperability solutions typically rely on third-party service providers to assist in completion, which brings about security risks such as new trust assumptions and single-point-of-failure issues. Completing interoperability among Rollups based on the native chain does not require introducing new trust assumptions, but will consume the computing and storage resources of the native chain, reduce the transaction throughput of the native chain, and thus seriously affect the performance of cross-Rollup. Based on this, this study proposes a cross-Rollup mechanism based on a native blockchain. By aggregating and processing transactions in batches, it effectively reduces the on-chain average computation and storage resource costs of individual transactions. Specifically, a transaction validity proof scheme based on zero-knowledge proof is proposed to significantly reduce the on-chain computation overhead of transaction validity verification. A transaction storage scheme based on index table data compression is proposed, reducing the average on-chain storage overhead of cross-Rollup transactions. An aggregation scale balance adjustment algorithm is proposed, which obtains the optimal aggregation scale, achieving a balance between on-chain resource consumption and processing latency. Finally, this study validates the proposed solution through experiments. The experimental results demonstrate that under the condition of complete trustlessness, the proposed solution reduces on-chain computing and storage overheads while achieving a balance between on-chain resource consumption and processing latency. Moreover, compared to existing cross-Rollup solutions, the proposed solution exhibits good system throughput.