diff --git a/include/gp/ipc/bottleneck.hpp b/include/gp/ipc/bottleneck.hpp
new file mode 100644
index 0000000..8928893
--- /dev/null
+++ b/include/gp/ipc/bottleneck.hpp
@@ -0,0 +1,53 @@
+#pragma once
+
+#include <atomic>
+
+#include "gp_config.hpp"
+
+namespace gp {
+    template<typename T>
+    class lock_guard {
+        T& ref;
+    public:
+        lock_guard(T& _ref)
+        : ref(_ref)
+        {
+            ref.lock();
+        }
+
+        ~lock_guard() {
+            ref.unlock();
+        }
+    };
+
+    /**
+     * @brief A fast mutual exclusion handler WITHOUT deadlock detection
+     */
+    class fast_bottleneck {
+        std::atomic_bool flag;
+    public:
+        fast_bottleneck() = default;
+        fast_bottleneck(fast_bottleneck&) = delete;
+        fast_bottleneck(fast_bottleneck&&) = delete;
+
+        void lock() {
+            while(not try_lock());
+        }
+
+        [[nodiscard]] bool try_lock() {
+            bool f = false;
+            bool t = true;
+            return flag.compare_exchange_strong(f,t,std::memory_order::acquire);
+        }
+
+        void unlock() {
+            gp_config::assertion(try_unlock(), "Unlocking failed in fast_bottleneck: potential double unlocking issue");
+        }
+
+        [[nodiscard]] bool try_unlock() {
+            bool f = false;
+            bool t = true;
+            return flag.compare_exchange_strong(t,f,std::memory_order::release);
+        }
+    };
+}
\ No newline at end of file
diff --git a/include/gp/system/scheduling/simple_scheduling.hpp b/include/gp/system/scheduling/simple_scheduling.hpp
index bd5abee..dc4a853 100644
--- a/include/gp/system/scheduling/simple_scheduling.hpp
+++ b/include/gp/system/scheduling/simple_scheduling.hpp
@@ -1,43 +1,33 @@
 #pragma once
-#include "gp/system/system.hpp"
 
-// TODO: implement unsafe_bottleneck to use in the locked list
+#include "gp/ipc/bottleneck.hpp"
+#include "gp/system/system.hpp"
 
 namespace gp{
 	namespace system {
 		class simple_scheduling : public gp::system::scheduling_scheme {
 			struct locked_circular_buffer {
-				std::atomic_bool lock;
 				gp::array<gp::system::task_queue::node_ptr, gp_config::limits::max_processes> processes;
 				size_t read_idx = 0;
 				size_t write_idx = 0;
+				fast_bottleneck lock;
 
 				void push(gp::system::task_queue::node_ptr node) {
-					bool t = true;
-					bool f = false;
-					while(not lock.compare_exchange_strong(f,t)){}
-					{
-						gp_config::assertion((write_idx + 1)%processes.size() != read_idx, "bad push to circular buffer");
-						processes[write_idx] = node;
-						write_idx=(write_idx+1)%processes.size();
-					}
-					while(not lock.compare_exchange_strong(t,f)){}
+					auto guard = lock_guard(lock);
+					gp_config::assertion((write_idx + 1)%processes.size() != read_idx, "bad push to circular buffer");
+					processes[write_idx] = node;
+					write_idx=(write_idx+1)%processes.size();
 				}
 
 				gp::system::task_queue::node_ptr pop() {
-					bool t = true;
-					bool f = false;
 					gp::system::task_queue::node_ptr ret;
-					while(not lock.compare_exchange_strong(f,t)){}
-					{
-						if(read_idx == write_idx) {
-							ret = nullptr;
-						} else {
-							ret = processes[read_idx];
-							read_idx=(read_idx+1)%processes.size();
-						}
+					auto guard = lock_guard(lock);
+					if(read_idx == write_idx) {
+						ret = nullptr;
+					} else {
+						ret = processes[read_idx];
+						read_idx=(read_idx+1)%processes.size();
 					}
-					while(not lock.compare_exchange_strong(t,f)){}
 					return ret;
 				}
 				
diff --git a/include/gp/utils/allocators/bottleneck_allocator.hpp b/include/gp/utils/allocators/bottleneck_allocator.hpp
new file mode 100644
index 0000000..80cc44a
--- /dev/null
+++ b/include/gp/utils/allocators/bottleneck_allocator.hpp
@@ -0,0 +1,58 @@
+#pragma once
+
+#include "gp/utils/allocators/allocator.hpp"
+#include "gp/ipc/bottleneck.hpp"
+
+namespace gp {
+    class bottleneck_allocator_front : public allocator {
+        allocator& backend;
+        fast_bottleneck lock;
+    public: 
+        bottleneck_allocator_front(allocator& _backend) 
+        : backend(_backend)
+        , lock()
+        {}
+
+		/**
+		 * @brief Allocates memory, THREAD SAFE
+		 *
+		 * @param sz the amount of bytes to allocate
+		 * 
+		 * @return the allocated memory as a pointer on success
+		 * @return nullptr if it failed allocating
+		 */
+		virtual void* allocate(size_t sz) {
+            auto guard = lock_guard(lock);
+            return backend.allocate(sz);
+        }
+
+		/**
+		 * @brief Deallocates memory, THREAD SAFE
+		 *
+		 * @param ptr the memory to deallocate
+		 * 
+		 * @return true if the memory was successfully deallocated
+		 * @return false if the memory was not deallocated
+		 */
+		virtual bool deallocate(void* ptr) {
+            auto guard = lock_guard(lock);
+            return backend.deallocate(ptr);
+        }
+
+		/**
+		 * @brief Tries to reallocate memory, THREAD SAFE
+		 *
+		 * @param ptr The memory to reallocate
+		 * @param sz The new size we want to give the memory
+		 * 
+		 * @return true if reallocation was successful
+		 * @return false if the reallocation failed
+		 */
+		virtual bool try_reallocate(void* ptr, size_t sz) {
+            auto guard = lock_guard(lock);
+            return backend.try_reallocate(ptr, sz);
+        }
+
+		virtual ~bottleneck_allocator_front() = default;
+    };
+}
\ No newline at end of file