This paper investigates the performance of a MIMO broadcast channel with finite rate feedback, where each receiver has perfect channel knowledge but the transmitter only receives quantized information about the channel. The focus is on the zero-forcing transmission technique, and the key findings are: 1. **Throughput Degradation**: The throughput of a feedback-based zero-forcing system is bounded if the SNR is taken to infinity and the number of feedback bits per mobile is kept fixed. 2. **Feedback Rate Scaling**: To achieve the full multiplexing gain of \(M\), the feedback rate per mobile must be increased linearly with the SNR (in dB) at the rate: \[ B = (M - 1) \log_2 P \approx \frac{M - 1}{3} P_{dB} \] 3. **Multiplexing Gain**: Scaling the number of feedback bits according to \(B = \alpha \log_2 P\) for any \(\alpha < M - 1\) results in a strictly inferior multiplexing gain of \(M \left(\frac{\alpha}{M-1}\right)\). 4. **Random Vector Quantization (RVQ)**: The use of RVQ, which is a simple and effective quantization method, is analyzed and shown to perform close to optimal quantization. The paper also discusses the implications of these findings for system design, highlighting that the feedback requirements are significantly higher in downlink channels compared to point-to-point MIMO channels.This paper investigates the performance of a MIMO broadcast channel with finite rate feedback, where each receiver has perfect channel knowledge but the transmitter only receives quantized information about the channel. The focus is on the zero-forcing transmission technique, and the key findings are: 1. **Throughput Degradation**: The throughput of a feedback-based zero-forcing system is bounded if the SNR is taken to infinity and the number of feedback bits per mobile is kept fixed. 2. **Feedback Rate Scaling**: To achieve the full multiplexing gain of \(M\), the feedback rate per mobile must be increased linearly with the SNR (in dB) at the rate: \[ B = (M - 1) \log_2 P \approx \frac{M - 1}{3} P_{dB} \] 3. **Multiplexing Gain**: Scaling the number of feedback bits according to \(B = \alpha \log_2 P\) for any \(\alpha < M - 1\) results in a strictly inferior multiplexing gain of \(M \left(\frac{\alpha}{M-1}\right)\). 4. **Random Vector Quantization (RVQ)**: The use of RVQ, which is a simple and effective quantization method, is analyzed and shown to perform close to optimal quantization. The paper also discusses the implications of these findings for system design, highlighting that the feedback requirements are significantly higher in downlink channels compared to point-to-point MIMO channels.