A note on the unitarization of the multiperipheral model