Lincoln Thurlow presents two talks on Erasure Codes and Active Networks Talk 1 (15-20 Minutes) Erasure Codes [Professor Miller]: Lincoln will present a re-cap of last Spring quarter’s work on Erasure codes using AVX2 Intel intrinsic with updated performance metrics not presented at our CRSS/SSRC Spring Retreat. Our most recent results show that AVX2 instructions enhance the performance Galois Field multiplication of a constant by a region for both GF(2^8) and GF(2^16) fields. Our results vary by region multiplication size based on the size of the L1, L2, and L3 cache sizes. This concurs with previous Erasure and SIMD literature showing that cache size can be and is a bottleneck. Our AVX2 implementation reaches peak speeds of 18 GB/s for 100KB region sizes in GF(2^8) and 9 GB/s for 100KB region sizes for GF(2^16) fields. These peak speeds show a 30% increase in performance over previous constant by region multiplication implementations using SSSE3. Talk 2 (20-30 Minutes) Active Networks [Professor Obraczka, Associate Adjunct Professor Smith]: Lincoln will discuss research in Active Networks from the late 1990’s and early 2000’s that introduced the concept of executable code within the network. Due to inherent risks of executable code in the network with respect to security and performance as well as a lack of a killer application to motivate future research, active networking funding dried up and research in the area halted. Looking back at active networking research and applying new technologies from hardware and software virtualization to more sophisticated software routing architectures provides the necessary foundation for developing a secure and fast practical active networking architecture. Our architecture SCAN proposes a language agnostic framework for current and future networking architectures. SCAN’s provides a more expressive API for networking programmability and addresses control plane decentralization problems with the current Software-Defined Networking architecture.