Re: [mat-wg] RIR prefix size distribution in RIPEstat