diff --git a/enhancements/20221213-reimplement-longhorn-engine-with-SPDK.md b/enhancements/20221213-reimplement-longhorn-engine-with-SPDK.md
index b805ade..3dca29d 100644
--- a/enhancements/20221213-reimplement-longhorn-engine-with-SPDK.md
+++ b/enhancements/20221213-reimplement-longhorn-engine-with-SPDK.md
@@ -36,7 +36,7 @@ These are the reasons that have driven us:
 
 SPDK implements a JSON-RPC 2.0 server to allow external management tools to dynamically configure SPDK components ([documentation](https://spdk.io/doc/jsonrpc.html)).
 
-What we aim is to create an external orchestrator that, with JSON-RPC calls towards multiple instances of `spdk_tgt` app running in different machines, could manage the durability and reliability of data. Actually, not all needed functionalities to do that are already available in SPDK, so some new JSON-RPC commands will be developed over SPDK. The name of such an orchestrator could be `longhorn-spdk-agent`, taking as example [Storage Management Agent](https://spdk.io/doc/sma.html) that has some similar functionalities, but `longhorn-spdk-engine` is a best fit to make better understand its continuity with actul `longhorn-engine`.
+What we aim is to create an external orchestrator that, with JSON-RPC calls towards multiple instances of `spdk_tgt` app running in different machines, could manage the durability and reliability of data. Actually, not all needed functionalities to do that are already available in SPDK, so some new JSON-RPC commands will be developed over SPDK. The name of such an orchestrator could be `longhorn-spdk-agent`, taking as example [Storage Management Agent](https://spdk.io/doc/sma.html) that has some similar functionalities, but `longhorn-spdk-engine` is a best fit to make better understand its continuity with actual `longhorn-engine`.
 
 * The main purpose of `longhorn-spdk-engine` is to create and export via NVMe-oF logical volumes from multiple replica nodes (one of them likely local), attach to these volumes on a controller node, use resulting bdevs to create a RAID1 bdev and exporting it via NVMe-oF locally. At this point NVMe Kernel module can be used to connect to this NVMe-oF subsystem and so to create a block device `/dev/nvmeXnY` to be used by the Longhorn CSI driver. In this way we will have multiple replica of the same data written on this block device.
 * Below a diagram that shows the control plane of the proposal ![SPDK New Architecture](./image/spdk-control-plane.png)