A 6 F800 node Isilon cluster has one subnet with two IP pools with LACP aggregated 10GigE interfaces from all the nodes in the 'Data-Zone' Access Zone. The 'NFS-Pool' with dynamic IP Allocation…

A 6 F800 node Isilon cluster has one subnet with two IP pools with LACP aggregated 10GigE interfaces from all the nodes in the 'Data-Zone' Access Zone. The 'NFS-Pool' with dynamic IP Allocation schema and 'SMBPool' with static IP Allocation schema are being used. All the Linux clients are accessing Isilon exports using NFSv3 via the 'NFS_Pool'. All the Windows clients are accessing Isilon shares using SMBv3 via the 'SMB_Pool'. While deploying Isilon, the customer reserves the whole network subnet for Isilon. The customer is facing performance issues with a new workload that is accessing the Isilon via Kerberized NFSv4. What is recommended when creating a new IP pool for NFSv4 clients?