The Other Doors, Named
Questi contenuti non sono ancora disponibili nella tua lingua.
Module 5 in the chapter The Doors In, chapter 4 of the Data 360 Architect course. Each module stands on its own. Previous: The Push Door: the Ingestion API
The question this module answers: What are the other ways into Data 360, and where does this course teach each one?
We’ve walked through two doors closely: Amazon S3, where Data 360 pulls, and the Ingestion API, where you push. There are a lot more. This module names the rest, one at a time, so you know they exist and where to find them. You don’t need to master them all today. You need to recognize them when a source shows up.
The good news: almost every one works the same way. A connection is set up once and holds the credentials (module 2). A data stream on that connection brings the data into a data lake object. And then you map it. Once you’ve done S3, you’ve seen the pattern.
Salesforce’s own clouds
Section titled “Salesforce’s own clouds”These are pulls, through connectors Salesforce builds and maintains.
- The Salesforce CRM connector brings in your orgs, one or many. It has its own decisions (which objects, which fields, upsert or full refresh), so it gets a whole chapter: chapter 5.
- Marketing Cloud and B2C Commerce each have their own connector.
- Marketing Cloud Personalization, what I knew as Interaction Studio, has a connector for behavior on websites and apps.
If the data is already in one of these, you don’t build anything. You set up the connection and pick what to bring in.
Files: cloud storage and SFTP
Section titled “Files: cloud storage and SFTP”- Amazon S3, Microsoft Azure Blob Storage and Google Cloud Storage are the same door, on three clouds. Data 360 pulls files on a schedule, and every decision from module 3 applies: which files, the refresh mode, and how often.
- SFTP. You can take your CSV files and push them to an SFTP server, a secure file transfer server. Salesforce’s page says it carries “CSV and Parquet data files from your SFTP server into Data 360 using SSH.” Notice what that means. From your side, it’s a push: your job drops the file on the server. From Data 360’s side, it’s a pull: the connection, with its SSH key, collects the file. So it behaves like S3. It’s the door for systems that can only export files, and Salesforce lists marketing applications such as Adobe Campaign and Oracle Responsys among them.
Through code or middleware
Section titled “Through code or middleware”- MuleSoft comes in through the Ingestion API. Salesforce says the same thing: set up the Ingestion API first, “and then Mulesoft Anypoint Connector.” So everything in module 4 still applies: the schema, streaming or bulk, the limits. MuleSoft adds the connections on the other side, to systems like SAP or a legacy database, without writing your own code for each.
- Your own server code calling the Ingestion API is module 4. For a server signing in with no person present, Salesforce’s architects point to a connected app with OAuth 2.0, for example the JWT bearer flow. The APIs, in and out, are chapter 14.
As it happens: websites, apps and event streams
Section titled “As it happens: websites, apps and event streams”- The web and mobile SDKs. You can set up your websites and mobile apps to send behavior in as people use them: a page viewed, an item added to a cart. On a website it’s a JavaScript script on each page; in an app it’s a native SDK. Salesforce’s architects flag the catch: this “depends on client-side SDKs,” so someone owns the script, its versions and its event schema.
- Amazon Kinesis and Amazon MSK (Kafka). If your company already runs an event platform, Data 360 can read from it directly, for high-volume operational or device events.
Both are chapter 13, real time.
Don’t move it: zero copy
Section titled “Don’t move it: zero copy”When I first recorded these doors, I hadn’t found a way to bring Snowflake data in except through one of the other doors. That’s changed. With inbound zero copy, Data 360 queries data where it lives, in platforms such as Snowflake and Google BigQuery. Nothing is copied in. It’s the door you don’t walk through. Chapter 15.
Documents, not rows: unstructured sources
Section titled “Documents, not rows: unstructured sources”Everything so far is rows and columns. Data 360 can also bring in documents: files, knowledge articles, web pages, content from places like SharePoint or Google Drive. They land in an unstructured data lake object, and they’re used for search and for grounding AI answers. Different shape, different purpose, its own chapter: chapter 18.
And hundreds more
Section titled “And hundreds more”Salesforce lists over 270 connectors (module 2). Before anyone writes integration code, look at that list. The source may already have a connector.
And look at the method column while you’re there. On Salesforce’s connector page in November 2026, 115 of the 184 connectors listed offer zero copy as well as batch. So for many sources, the first question at setup isn’t only “which connector?” It’s “copy the data in, or read it where it is?” Chapter 15 helps you answer it.
Two more places to look before anyone builds. In your org’s list of sources, Installed Data Kits & Packages brings in data streams that come already set up in a data kit or a package (you saw the tile in module 3’s screenshot; data kits are chapter 22). And Salesforce’s marketplace, AgentExchange, has Data 360 solutions: in Salesforce’s words, “easy-to-install apps or agents that help you activate, extend, enhance, and enrich your Data 360 implementations.” They come in three groups: data enrichment apps, data activation apps, and consultants who specialize in Data 360. Treat an app like any partner that touches your data: know what it does with it (chapter 3, module 5).
And when none fits, you still have the two general doors. As I said in the video, with the cloud storage or the Ingestion API “you can use almost any kind of data and push through there.”
Four ways in, by how soon
Section titled “Four ways in, by how soon”There’s another way to sort all of these doors: by how soon the data has to be there. Four ways in:
- Batch and streaming: planned transfers, on a schedule or as changes come. Most of this chapter.
- Unstructured: documents, ready for search a few minutes after they land. Salesforce’s own workshop says it “will take a few minutes for the data to be ingested.”
- Real time: for one person, right now. Salesforce’s documentation: “95% of events complete end-to-end ingestion from capture through real-time layer activation within approximately 500 milliseconds under the Sub-Second Real-Time Profile.” Chapter 13.
- Zero copy: not ingested at all; read where it lives. Chapter 15.
My advice: not everything needs to be ingested, streamed, real time or zero copy. Pick by what the business needs from the data, and by what it costs (chapter 21).
All the doors, in one table
Section titled “All the doors, in one table”| Door | Pull or push | Where it’s taught |
|---|---|---|
| Salesforce CRM connector | Pull | Chapter 5 |
| Marketing Cloud, B2C Commerce, Marketing Cloud Personalization | Pull | Named here |
| File Upload, a CSV from your own computer | You upload it by hand; no schedule | Chapter 4, module 3 |
| Amazon S3, Azure Blob Storage, Google Cloud Storage | Pull | Chapter 4, module 3 |
| SFTP | You push to the server; Data 360 pulls from it | Named here |
| Ingestion API, streaming and bulk | Push | Chapter 4, module 4 |
| MuleSoft | Push, through the Ingestion API | Named here; module 4 applies |
| Web and mobile SDKs | Push | Chapter 13 |
| Amazon Kinesis and MSK | Pull, from your event platform | Chapter 13 |
| Inbound zero copy | Neither: queried in place | Chapter 15 |
| Unstructured sources | Mostly pull; cloud storage can push a notice when a file changes, and a few connectors are push | Chapter 18 |
What an architect decides here
Section titled “What an architect decides here”- For each source: is there already a connector? Check the list before anyone builds.
- When a system can only export files: cloud storage or SFTP, whichever the source team can run.
- When middleware is in the picture: it still pushes through the Ingestion API, so module 4’s limits are your limits.
Explore
Section titled “Explore”All checked 2026-11-14.
- Data 360 Integration Patterns (opens in a new tab), Salesforce Architects: every inbound pattern.
- Secure File Transfer Protocol (SFTP) Structured Connector (opens in a new tab), Salesforce Developers.
- MuleSoft Anypoint Connector for Salesforce Data 360 (opens in a new tab), Salesforce Developers.
- Data 360 Connectors and Integrations (opens in a new tab), Salesforce Developers: the full list, with each connector’s method (counted November 2026).
- Real-Time Ingestion (opens in a new tab), Salesforce Developers, checked 2026-11-14: the 500 millisecond figure and its limits.
- Exercise 1: Ingest Unstructured Data (opens in a new tab), Salesforce Developers workshop, checked 2026-11-14: documents from S3, with file notifications.
- Data 360 Solutions on AgentExchange: Quick Look (opens in a new tab), Trailhead, checked 2026-11-14: what Data 360 solutions are, in three groups.