While cloud computing provides high efficiency calculation and massive storage for workflow application

it also brings challenges to the protection of private data such as trade secrets. Because protecting private data will result in an increase in data transmission time

a new data distribution algorithm based on data collaboration failure degree is proposed to improve the existing data allocation strategies. Relying on private cloud safety and public cloud computer capacity in the initial stage

the non-privacy data is divided by the static allocation method to improve the initial data

and in the running stage

the generated data is constantly adjusted to get the data distribution plan. Experiments show that the improved method is effective in reducing the data transmission time.