In my earlier blog (Junos High Availability Design Guide) I discussed how to make use of redundant routing engines by configuring features like (GRES, NSR, NSB) for reducing downtime to the minimum possible level.
The real problem is that one RE is active at one time and all PFEs must be connected with the active RE. In case of failure of the primary Routing Engine (RE), the backup RE will take over, and all PFEs will now need to connect to the new primary RE. This scenario can cause momentary disruption of services.
MC-LAG (Active-Active) is the correct solution to above described problem as it offers 2 active REs in 2 different devices/ chassis. Read the full solution in my blog here.