You only need one capacity to drive all your Microsoft Fabric experiences, including Microsoft OneLake. Keep reading if you want a detailed example of how OneLake consumes storage and compute.
Overview
OneLake comes automatically with every Fabric tenant and is designed to be the single place for all your analytics data. All the Fabric data items are prewired to store data in OneLake. For example, when you store data in a lakehouse or warehouse, your data is natively stored in OneLake.
With OneLake, you pay for the data stored, similar to services like Azure Data Lake Storage (ADLS) Gen2 or Amazon S3. However, unlike other services, OneLake doesn't include a separate charge for transactions (for example, reads, writes) to your data. Instead, transactions consume from existing Fabric capacity that is also used to run your other Fabric experiences. For information about pricing, which is comparable to ADLS Gen2, see Fabric pricing.
To illustrate, let’s walk through an example.
Let’s say you purchase an F2 SKU with 2 Capacity Units (CU) every second. Let’s name this Capacity1.
You then create Workspace1 and upload a 450 MB file to a lakehouse using the Fabric portal. This action consumes both OneLake storage and OneLake transactions.
Now, let’s dive into each of these dimensions.
OneLake Storage
Since OneLake storage operates on a pay-as-you-go model, a separate charge for "OneLake Storage" appears in your bill corresponding to the 450 MB of data stored.
If you're a capacity admin, you can view your storage consumption in the Fabric Capacity Metrics app. Open the Storage tab and choose Experience as lake to see the cost of OneLake storage. If you have multiple workspaces in the capacity, you can see the storage per workspace.
下圖顯示兩個資料列：計費記憶體 和 目前記憶體。 可計費記憶體會反映當月累計數據使用量。 因為儲存資料的總費用不是在某一天收取，而是在整個月中按比例分攤。 您可以將每月價格估計為可計費記憶體 （GB） 乘以每個 GB 每月的價格。
例如，在第 1 天儲存 1 TB 的數據，會新增至 33 GB 的每日計費記憶體。 第一天是 1 TB / 30 天 = 33 GB，而且每天會新增 33 GB，直到當月結束為止。 OneLake soft delete protects individual files from accidental deletion by retaining files for a default retention period before it's permanently deleted. Soft-deleted data is billed at the same rate as active data.
OneLake Compute
對 OneLake 的要求（例如讀取、寫入或清單）會耗用 Fabric 容量。 OneLake maps APIs to operations like ADLS. 容量計量應用程式中會顯示每個作業的容量使用量。 在上述範例中，檔案上傳導致寫入交易耗用 127.46 CU 秒。 This consumption is reported as OneLake Write via Proxy under the operation name column in the capacity metrics App.
Now if you read this data using a notebook. You consume 1.39 CU seconds of read transactions. 此耗用量在計量應用程式中會以 OneLake Read via Redirect 的形式回報。 請參閱 OneLake 耗用量頁面，以瞭解每種作業類型如何取用容量單位。
若要深入瞭解計量應用程式的各種術語，請參閱 瞭解計量應用程式計算頁面 - Microsoft Fabric。
You may be wondering, how do shortcuts affect my OneLake usage? In the above example, both storage and compute are billed to Capacity1. Now, let’s say you have a second capacity Capacity2, that contains Workspace2. You create a lakehouse and create a shortcut to the parquet file you uploaded in Workspace1. You create a notebook to query the parquet file. As Capacity2 accesses the data, the compute or transaction cost for this read operation consumes CU from Capacity2. The storage continues to be billed to Capacity1.
如果 Capacity2 已暫停，但 Capacity1 為使用中，您無法透過 Workspace2 （Capacity2） 中的快捷方式讀取數據，但可以直接在 Workspace1 中存取數據（Capacity1）。
如果 Capacity1 已暫停且 Capacity2 為使用中，則您無法讀取 Workspace1 （Capacity1） 中的數據，但您仍然可以使用 Workspace2 中的快捷方式來使用數據。 在這兩種情況下，由於數據仍儲存在 Capacity1 中，記憶體成本仍會計入 Capacity1
If your CU consumption exceeds the capacity limit, throttling may occur, causing transactions to be delayed or rejected temporarily.
啟動 Fabric 的 60 天免費試用，以探索 OneLake 和其他功能，並訪問 Fabric 論壇 以提出問題。