]> Transloadit

marius@transloadit.com

Apple Inc.

guoye_zhang@apple.com

Cloudflare

lucas@lucaspardue.com

Web and Internet Transport HTTP HTTP, upload, resumable Data transfer using the Hypertext Transfer Protocol (HTTP) is often interrupted by canceled requests or dropped connections. If the intended recipient can indicate how much of the data was received prior to interruption, a sender can resume data transfer at that point instead of attempting to transfer all of the data again. HTTP range requests support this concept of resumable downloads from server to client. This document describes a mechanism that supports resumable uploads from client to server using HTTP. About This Document Status information for this document may be found at . Discussion of this document takes place on the HTTP Working Group mailing list (), which is archived at . Working Group information can be found at . Source for this draft and an issue tracker can be found at .

Introduction Data transfer using the Hypertext Transfer Protocol () is often interrupted by canceled requests or dropped connections. If the intended recipient can indicate how much of the data was received prior to interruption, a sender can resume data transfer at that point instead of attempting to transfer all of the data again. HTTP range requests (see ) support this concept of resumable data transfers for downloads from server to client. This specification defines a mechanism for resumable uploads from client to server in a way that is backwards-compatible with conventional HTTP uploads. When an upload is interrupted, clients can send subsequent requests to query the server state and use this information to send the remaining representation data. Alternatively, they can cancel the upload entirely. Unlike ranged downloads, this protocol does not support transferring an upload as multiple requests in parallel. Utilizing resumable uploads, applications can recover from unintended interruptions, but also interrupt an upload on purpose to later resume it, for example, when a user wants to pause an upload, the device's network connectivity changes, or bandwidth should be saved for higher priority tasks. The document introduces the concept of an upload resource to facilitate resumable uploads () and defines new header fields to communicate the state of the upload (), the status code 104 (Upload Resumption Supported) to indicate the server's support for resumable uploads (), and the application/partial-upload media type to label partial representation data when resuming an upload ().

Conventions and Definitions The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 when, and only when, they appear in all capitals, as shown here. Some examples in this document contain long lines that may be folded, as described in . The terms Structured Header, Item, Dictionary, String, Integer, and Boolean are imported from . The terms "representation", "representation data", "representation metadata", "content", "client" and "server" are from . The term "URI" is used as defined in . The term "patch document" is taken from . An upload resource is a temporary resource on the server that facilitates the resumable upload of one representation ().

Overview Resumable uploads are supported in HTTP through use of a temporary resource, an upload resource (), that is separate from the resource being uploaded to and specific to that upload. By interacting with the upload resource, a client can retrieve the current offset of the upload (), append to the upload (), and cancel the upload (). The remainder of this section uses examples to illustrate different interactions with the upload resource. HTTP message exchanges, and thereby resumable uploads, use representation data (see ). This means that resumable uploads can be used with many forms of content, such as static files, in-memory buffers, data from streaming sources, or on-demand generated data.

Example 1: Complete upload of representation data with known size In this example, the client first attempts to upload representation data with a known size in a single HTTP request to the resource at /project/123/files. An interruption occurs and the client then attempts to resume the upload using subsequent HTTP requests to the upload resource at /uploads/abc. 1) The client notifies the server that it wants to begin an upload (). The server reserves the required resources to accept the upload from the client, and the client begins transferring the entire representation data in the request content. An interim response can be sent to the client, which signals the server's support of resumable upload as well as the upload resource's URI via the Location header field ().

Upload Creation | | | | | Reserve resources | | for upload | |-----------------. | | | | |<----------------' | | | 104 Upload Resumption Supported | | Location: /uploads/abc | |<-------------------------------------------| | | X--------------Flow Interrupted--------------X ]]>

2) If the connection to the server is interrupted, the client might want to resume the upload. However, before this is possible the client needs to know the amount of representation data that the server received before the interruption. It does so by retrieving the offset () from the upload resource.

Offset Retrieval | | | | 204 No Content | | Upload-Offset: X | |<------------------------------------------------| | | ]]>

3) The client can resume the upload by sending the remaining representation data to the upload resource (), appending to the already stored representation data in the upload using the application/partial-upload media type. The Upload-Offset value is included to ensure that the client and server agree on the offset that the upload resumes from. Once the remaining representation data is transferred, the server processes the entire representation and responds with whatever the initial request to /project/123/files would have produced if it had not been interrupted, e.g. a 200 (OK) response.

Upload Append | | | | 200 OK | |<------------------------------------------------| | | ]]>

4) If the client is not interested in completing the upload, it can instruct the upload resource to delete the upload and free all related resources ().

Upload Cancellation | | | | 204 No Content | |<------------------------------------------------| | | ]]>

Example 2: Upload as a series of parts In some cases, clients might prefer to upload a representation as a series of parts sent serially across multiple HTTP messages. One use case is to overcome server limits on HTTP message content size. Another use case is where the client does not know the final size of the representation data, such as when the data originates from a streaming source. This example shows how the client, with prior knowledge about the server's resumable upload support, can upload parts of a representation incrementally. 1) If the client is aware that the server supports resumable upload, it can start an upload with the Upload-Complete field value set to false and the first part of the representation.

Upload creation with partial representation data | | | | 201 Created | | Location: /uploads/abc | |<------------------------------------------------| | | ]]>

2) Subsequent, intermediate parts are appended () with the Upload-Complete field value set to false, indicating that they are not the last part of the representation data. The offset value in the Upload-Offset header field is taken from the previous response when creating the upload or appending to it.

Appending partial representation data to upload | | | | 204 No Content | |<------------------------------------------------| | | ]]>

3) If the connection was interrupted, the client might want to resume the upload, similar to the previous example (). The client retrieves the offset () to learn the amount of representation data received by the server and then continues appending the remaining parts to the upload as in the previous step.

Resuming an interrupted upload | | | | 204 No Content | | Upload-Offset: Y | |<------------------------------------------------| | | | PATCH /uploads/abc | | Upload-Complete: ?0 | | Upload-Offset: Y | | Content-Type: application/partial-upload | |------------------------------------------------>| | | | 204 No Content | |<------------------------------------------------| | | ]]>

4) The request to append the last part of the representation data has a Upload-Complete field value set to true to indicate the complete transfer. Once the remaining representation data is transferred, the server processes the entire representation and responds with whatever the initial request to /project/123/files would have produced if it had received the entire representation, e.g. a 200 (OK) response.

Appending remaining representation data | | | | 200 OK | |<------------------------------------------------| | | ]]>

Upload Resource A resumable upload is enabled through interaction with an upload resource. When a resumable upload begins, the server is asked to create an upload resource through a request to another resource (). This upload resource is responsible for handling the upload of a representation. Using the upload resource, the client can query the upload progress (), append representation data (), or cancel the upload (). An upload resource is specific to the upload of one representation. For uploading multiple representations, multiple upload resources have to be used. The server can clean up an upload resource and make it inaccessible immediately after the upload is complete. However, if a client didn't receive the last response acknowledging the upload's completion and the upload resource is not available anymore, the client cannot verify the upload's state with the server. Therefore, the server SHOULD keep the upload resource available for a reasonable amount of time after the upload is complete. An upload resource SHOULD NOT reuse the URI from a previous upload resource, unless reasonable time has passed to ensure that no client will attempt to access the previous upload resource. Otherwise, a client might access the upload resource corresponding to a different representation than it intends to transfer.

State The state of an upload consists of the following properties that are tracked by the upload resource.

Offset The offset is the number of bytes from the representation data that have been received, either during the creation of the upload resource () and by appending to it (). The offset can be retrieved from the upload resource () and is required when appending representation data () to synchronize the client and resource regarding the amount of transferred representation data. Representation data received by the upload resource cannot be removed again and, therefore, the offset MUST NOT decrease. If the upload resource loses representation data, the server MUST consider the upload resource invalid and reject further interaction with it. The Upload-Offset request and response header field conveys the offset. Upload-Offset is an Item Structured Header Field (). Its value is a non-negative Integer () and indicates the current offset as viewed by the message sender. Other values MUST cause the entire header field to be ignored. The Upload-Offset header field in responses serves as an acknowledgement of the received representation data and as a guarantee that no retransmission of it will be necessary. Clients can use this guarantee to free resources associated to transferred representation data.

Completeness An upload is incomplete until it is explicitly marked as completed by the client. After this point, no representation data can be appended anymore. The Upload-Complete request and response header field conveys the completeness state. Upload-Complete is an Item Structured Header Field (). Its value is a Boolean () and indicates whether the upload is complete or not. Other values MUST cause the entire header field to be ignored. An upload is marked as completed if a request for creating the upload resource () or appending to it () included the Upload-Complete header field with a true value and the request content was fully received.

Length The length of an upload is the number of bytes of representation data that the client intends to upload. Even the client might not know the total length of the representation data when starting the transfer, for example, because the representation is taken from a streaming source. However, a client SHOULD communicate the length to the upload resource as soon as it becomes known. There are two different ways for the client to indicate and the upload resource to discover the length from requests for creating the upload resource () or appending to it ():

• If the request includes the Upload-Complete field value set to true and a valid Content-Length header field, the request content is the remaining representation data. The length is then the sum of the current offset () and the Content-Length header field value.
• The request can include the Upload-Length request and response header field. Upload-Length is an Item Structured Header Field (). Its value is a non-negative Integer () and indicates the number of bytes of the entire representation data. Other values MUST cause the entire header field to be ignored.

If both indicators are present in the same request, their indicated lengths MUST match. If multiple requests include indicators, their indicated values MUST match. A server can use the problem type of "https://iana.org/assignments/http-problem-types#inconsistent-upload-length" () in responses to indicate inconsistent length values. The upload resource might not know the length until the upload is complete. Note that the length and offset values do not determine whether an upload is complete. Instead, the client uses the Upload-Complete () header field to indicate that a request completes the upload. The offset could match the length, but the upload can still be incomplete.

Limits An upload resource MAY enforce one or multiple limits, which are communicated to the client via the Upload-Limit response header field. Upload-Limit is a Dictionary Structured Header Field (). Its value is a Dictionary (). Other values MUST cause the entire header field to be ignored. The following key-value pairs are defined:

• The value under the max-size key specifies a maximum size for the representation data, counted in bytes. The server might not create an upload resource if the length () deduced from the upload creation request is larger than the maximum size. The upload resource can stop the upload if the offset () exceeds the maximum size. The value is an Integer.
• The value under the min-size key specifies a minimum size for the representation data, counted in bytes. The server might not create an upload resource if the length () deduced from the upload creation request is smaller than the minimum size or no length can be deduced at all. The value is an Integer.
• The value under the max-append-size key specifies a maximum size counted in bytes for the request content in a single upload append request (). The server might reject requests exceeding this limit. A client that is aware of this limit MUST NOT send larger upload append requests. The value is an Integer.
• The value under the min-append-size key specifies a minimum size counted in bytes for the request content in a single upload append request (). The server might reject requests below this limit. A client that is aware of this limit MUST NOT send smaller upload append requests. The value is an Integer. Requests completing the upload by including the Upload-Complete: ?1 header field are exempt from this limit.
• The value under the max-age key specifies the remaining lifetime of the upload resource in seconds counted from the generation of the response. After the resource's lifetime is reached, the server might make the upload resource inaccessible and a client SHOULD NOT attempt to access the upload resource as these requests will likely fail. The lifetime MAY be extended but SHOULD NOT be reduced unless the server makes the upload resource immediately inaccessible. The value is an Integer.

Except for the max-age limit, the existence of a limit or its value MUST NOT change throughout the lifetime of the upload resource. When parsing the Upload-Limit header field, unrecognized keys MUST be ignored and MUST NOT fail the parsing to facilitate the addition of new limits in the future. Keys with values other than defined here MUST be ignored. A server that supports the creation of a resumable upload resource () for a target URI MUST include the Upload-Limit header field with the corresponding limits in a response to an OPTIONS request sent to this target URI. If a server supports the creation of upload resources for any target URI, it SHOULD include the Upload-Limit header field with the corresponding limits in a response to an OPTIONS request with the * target unless the server is not capable of handling OPTIONS * requests. If the server does not apply any limits, it MUST use min-size=0 instead of an empty header value. A client can use an OPTIONS request to discover support for resumable uploads and potential limits before creating an upload resource. To reduce the liklihood of failing requests, the limits announced in an OPTIONS response SHOULD NOT be less restrictive than the limits applied to an upload once the upload resource has been created, unless the request to create an upload resource included additional information that warrants different limits. For example, a server might announce a general maximum size limit of 1GB, but reduce it to 100MB when the media type indicates an image.

Upload Creation

Client Behavior A client can start a resumable upload from any request that can carry content by including the Upload-Complete header field (). As a consequence, all request methods that allow content are possible, such as POST, PUT, and PATCH. The Upload-Complete header field is set to true if the request content includes the entire representation data that the client intends to upload. This is also a requirement for transparently upgrading to resumable uploads from traditional uploads (). If the client knows the representation data's length, it SHOULD include the Upload-Length header field () in the request to help the server allocate necessary resources for the upload and provide early feedback if the representation violates a limit (). The client SHOULD respect any limits () announced in the Upload-Limit header field in interim or final responses. In particular, if the allowed maximum size is less than the amount of representation data the client intends to upload, the client SHOULD stop the current request immediately and cancel the upload (). The request content can be empty. If the Upload-Complete header field is then set to true, the client intends to upload an empty representation. An Upload-Complete header field is set to false is also valid. This can be used to retrieve the upload resource's URI before transferring any representation data. Since interim responses are optional, this technique provides another mechanism to learn the URI, at the cost of an additional round-trip before data upload can commence. Representation metadata included in the initial request (see ) can affect how servers act on the uploaded representation data. The Content-Type header field () indicates the media type of the representation. The Content-Disposition header field () can be used to transmit a filename. The Content-Encoding header field () names the content codings applied to the representation. If the client received a final response with a

• 2xx (Successful) status code and the entire representation data was transferred in the request content, the upload is complete and the response belongs to the targeted resource processing the representation.
• 2xx (Successful) status code and not the entire representation data was transferred in the request content, the Location response header field points the client to the created upload resource. The client can continue appending representation data to it ().
• 4xx (Client Error) status code, the client SHOULD NOT attempt to retry or resume the upload, unless the semantics of the response allow or recommend the client to retry the request.
• 5xx (Server Error) status code or no final response at all due to connectivity issues, the client MAY automatically attempt upload resumption by retrieving the current offset () if it received the URI of the upload resource in a 104 (Upload Resumption Supported) interim response.

Server Behavior Upon receiving a request with the Upload-Complete header field, the server can choose to offer resumption support by creating an upload resource. If so, it SHOULD announce the upload resource by sending an interim response with the 104 (Upload Resumption Supported) status code and the Location header field pointing to the upload resource unless the server is not capable of sending interim responses. The interim response MUST include the Upload-Limit header field with the corresponding limits () if existing. The interim response allows the client to resume the upload even if the message exchange gets later interrupted. The resource targeted by this initial request is responsible for processing the representation data transferred in the resumable upload according to the method and header fields in the initial request, while the upload resource enables resuming the transfer. If the Upload-Complete request header field is set to true, the client intends to transfer the entire representation data in one request. If the request content was fully received, no resumable upload is needed and the resource proceeds to process the request and generate a response. If the Upload-Complete header field is set to false, the client intends to transfer the representation over multiple requests. If the request content was fully received, the server MUST announce the upload resource by referencing it in the Location response header field. Servers are RECOMMENDED to use the 201 (Created) status code. The response MUST include the Upload-Limit header field with the corresponding limits if existing. The server MUST record the length according to if the necessary header fields are included in the request. While the request content is being received, the server MAY send additional interim responses with a 104 (Upload Resumption Supported) status code and the Upload-Offset header field set to the current offset to inform the client about the upload progress. These interim responses MUST NOT include the Location header field. If the server does not receive the entire request content, for example because of canceled requests or dropped connections, it SHOULD append as much of the request content as possible to the upload resource. The upload resource MUST NOT be considered complete then.

Examples A) The following example shows an upload creation, where the entire 100 bytes are transferred in the initial request. The server sends multiple interim responses and one final response from processing the uploaded representation. B) The following example shows an upload creation, where only the first 25 bytes of a 100 bytes upload are transferred. The server acknowledges the received representation data and that the upload is not complete yet. The client can continue appending data. C) The following example shows an upload creation, where the server responds with a 5xx status code. Thanks to the interim response containing the upload resource URI, the client can resume the upload. D) The following example shows an upload creation being rejected by the server. The client cannot continue the upload.

Offset Retrieval

Client Behavior If the client wants to resume the upload after an interruption, it has to know the amount of representation data received by the upload resource so far. It can fetch the offset by sending a HEAD request to the upload resource. Upon a successful response, the client can continue the upload by appending representation data () starting at the offset indicated by the Upload-Offset response header field. The offset can be less than or equal to the number of bytes of representation data that the client has already sent. The client MAY reject an offset which is greater than the number of bytes it has already sent during this upload. The client is expected to handle backtracking of a reasonable length. If the offset is invalid for this upload, or if the client cannot backtrack to the offset and reproduce the same representation data it has already sent, the upload MUST be considered a failure. The client SHOULD cancel the upload () after rejecting the offset. The client MUST NOT perform offset retrieval while creation () or appending () is in progress as this can cause the previous request to be terminated by the server as described in . If the client received a response with a

• 2xx (Successful) status code, the client can continue appending representation data to it () if the upload is not complete yet.
• 307 (Temporary Redirect) or 308 (Permanent Redirect) status code, the client MAY retry retrieving the offset from the new URI.
• 4xx (Client Error) status code, the client SHOULD NOT attempt to retry or resume the upload, unless the semantics of the response allow or recommend the client to retry the request.
• 5xx (Server Error) status code or no final response at all due to connectivity issues, the client MAY retry retrieving the offset.

Server Behavior A successful response to a HEAD request against an upload resource

• MUST include the offset in the Upload-Offset header field (),
• MUST include the completeless state in the Upload-Complete header field (),
• MUST include the length in the Upload-Length header field if known (),
• MUST indicate the limits in the Upload-Limit header field (), and
• SHOULD include the Cache-Control header field with the value no-store to prevent HTTP caching ().

The resource SHOULD NOT generate a response with the 301 (Moved Permanently) and 302 (Found) status codes because clients might follow the redirect without preserving the HEAD method.

Example A) The following example shows an offset retrieval request. The server indicates the current offset and that the upload is not complete yet. The client can continue to append representation data. B) The following example shows on offset retrieval request for a completed upload. The client does not need to continue the upload.

Upload Append

Client Behavior A client can continue the upload and append representation data by sending a PATCH request with the application/partial-upload media type () to the upload resource. The request content is the representation data to append. The client MUST indicate the offset of the request content inside the representation data by including the Upload-Offset request header field. To ensure that the upload resource will accept request, the offset SHOULD be taken from an immediate previous response for retrieving the offset () or appending representation data (). The request MUST include the Upload-Complete header field. Its value is true if the end of the request content is the end of the representation data. If the content is then fully received by the upload resource, the upload will be complete. The request content can be empty. If the Upload-Complete field is then set to true, the client wants to complete the upload without appending additional representation data. If the client received a final response with a

• 2xx (Successful) status code and the remaining representation data was transferred in the request content, the upload is complete and the corresponding response belongs to the resource processing the representation according to the initial request (see ).
• 2xx (Successful) status code and not the entire remaining representation data was transferred in the request content, the client can continue appending representation data.
• 307 (Temporary Redirect) or 308 (Permanent Redirect) status code, the client MAY retry appending to the new URI.
• 4xx (Client Error) status code, the client SHOULD NOT attempt to retry or resume the upload, unless the semantics of the response allow or recommend the client to retry the request.
• 5xx (Server Error) status code or no final response at all due to connectivity issues, the client MAY automatically attempt upload resumption by retrieving the current offset ().

Server Behavior An upload resource applies a PATCH request with the application/partial-upload media type () by appending the patch document in the request content to the upload resource. If the upload resource does not receive the entire patch document, for example because of canceled requests or dropped connections, it SHOULD append as much of the patch document as possible, starting at its beginning and without discontinuities. Appending a continuous section starting at the patch document's beginning constitutes a successful PATCH as defined in . If the Upload-Offset request header field value does not match the current offset (), the upload resource MUST reject the request with a 409 (Conflict) status code. The response MUST include the correct offset in the Upload-Offset header field. The response can use the problem type of "https://iana.org/assignments/http-problem-types#mismatching-upload-offset" (). If the upload is already complete (), the server MUST NOT modify the upload resource and MUST reject the request. The server can use the problem type of "https://iana.org/assignments/http-problem-types#completed-upload" in the response (). If the Upload-Complete request header field is set to true, the client intends to transfer the remaining representation data in one request. If the request content was fully received, the upload is marked as complete and the upload resource SHOULD generate the response that matches what the resource, that was targeted by the initial upload creation (), would have generated if it had received the entire representation in the initial request. However, the response MUST include the Upload-Complete header field with a true value, allowing clients to identify whether a response, in particular error responses, is related to the resumable upload itself or the processing of the upload representation. If the Upload-Complete request header field is set to false, the client intends to transfer the remaining representation over multiple requests. If the request content was fully received, the upload resource acknowledges the appended data by sending a 2xx (Successful) response. If the request didn't complete the upload, any response, successful or not, MUST include the Upload-Complete header field with a false value, indicating that this response does not belong to the processing of the uploaded representation. The upload resource MUST record the length according to if the necessary header fields are included in the request. If the length is known, the upload resource MUST prevent the offset from exceeding the upload length by stopping to append bytes once the offset reaches the length, rejecting the request, marking the upload resource invalid and rejecting any further interaction with it. It is not sufficient to rely on the Content-Length header field for enforcement because the header field might not be present. While the request content is being received, the server SHOULD send interim responses with a 104 (Upload Resumption Supported) status code and the Upload-Offset header field set to the current offset to inform the client about the upload progress. These interim responses MUST NOT include the Location header field.

Example A) The following example shows an upload append request. The client transfers the next 100 bytes at an offset of 100 and does not indicate that the upload is then completed. The server generates one interim response and finally acknowledges the new offset: B) The next example shows an upload append, where the client transfers the remaining 200 bytes and completes the upload. The server processes the uploaded representation and generates the responding response, in this example containing extracted meta data:

Upload Cancellation

Client Behavior If the client wants to terminate the transfer without the ability to resume, it can send a DELETE request to the upload resource. Doing so is an indication that the client is no longer interested in continuing the upload, and that the server can release any resources associated with it.

Server Behavior Upon receiving a DELETE request, the server SHOULD deactivate the upload resource. The server SHOULD terminate any in-flight requests to the upload resource before sending the response by abruptly terminating their HTTP connection(s) or stream(s) as described in . The resource SHOULD NOT generate a response with the 301 (Moved Permanently) and 302 (Found) status codes because clients might follow the redirect without preserving the DELETE method.

Example The following example shows an upload cancellation:

Concurrency Resumable uploads, as defined in this document, do not permit uploading representation data in parallel to the same upload resource. The client MUST NOT perform multiple representation data transfers for the same upload resource in parallel. Even if the client is well behaving and doesn't send concurrent requests, network interruptions can occur in such a way that the client considers a request as failed while the server is unaware of the problem and considers the request still ongoing. The client might then try to resume the upload with the best intentions, resulting in concurrent requests from the server's perspective. Therefore, the server MUST take measures to prevent race conditions, data loss and corruption from concurrent requests to append representation data () and/or cancellation () to the same upload resource. In addition, the server MUST NOT send outdated information in responses when retrieving the offset (). This means that the offset sent by the server MUST be accepted in a subsequent request to append representation data if no other request to append representation data or cancel was received in the meantime. In other words, clients have to be able to use received offsets. The RECOMMENDED approach is as follows: If an upload resource receives a new request to retrieve the offset (), append representation data (), or cancel the upload () while a previous request for creating the upload () or appending representation data () is still ongoing, the resource SHOULD prevent race conditions, data loss, and corruption by terminating the previous request before processing the new request. Due to network delay and reordering, the resource might still be receiving representation data from an ongoing transfer for the same upload resource, which in the client's perspective has failed. Since the client is not allowed to perform multiple transfers in parallel, the upload resource can assume that the previous attempt has already failed. Therefore, the server MAY abruptly terminate the previous HTTP connection or stream. Since implementing this approach is not always technically possible or feasible, other measures can be considered as well. A simpler approach is that the server only processes a new request to retrieve the offset (), append representation data (), or cancellation () once all previous requests have been processed. This effectively implements exclusive access to the upload resource through an access lock. However, since network interruptions can occur in ways that cause the request to hang from the server's perspective, it might take the server significant time to realize the interruption and time out the request. During this period, the client will be unable to access the resource and resume the upload, causing friction for the end users. Therefore, the recommended approach is to terminate previous requests to enable quick resumption of uploads.

Status Code 104 (Upload Resumption Supported) The 104 (Upload Resumption Supported) status code is can be used for two purposes:

• When responding to requests to create uploads, an interim response with the 104 (Upload Resumption Supported) status code can be sent to indicate the server's support for resumable uploads, as well as the URI and limits of the corresponding upload resource in the Location and Upload-Limit header fields, respectively (see ). This notifies the client early about the ability to resume the upload in case of network interruptions.
• While processing the content of a request to append representation data or create an upload, the server can regularly send interim responses with the 104 (Upload Resumption Supported) status code to indicate the current upload progress in the Upload-Offset header field (see and ). This allows the client to show more accurate progress information about the amount of data receive by the server. In addition, clients can use this information to release representation data that was buffered, knowing that it doesn't have to be re-transmitted.

Media Type application/partial-upload The application/partial-upload media type describes a contiguous block from the representation data that should be uploaded to a resource. There is no minimum block size and the block might be empty. The block can be a subset of the representation data, where the start and/or end of the block don't line up with the start and/or end of the representation data respectively.

Problem Types

Mismatching Offset This section defines the "https://iana.org/assignments/http-problem-types#mismatching-upload-offset" problem type . A server can use this problem type when responding to an upload append request () to indicate that the Upload-Offset header field in the request does not match the upload resource's offset. Two problem type extension members are defined: the expected-offset and provided-offset members. A response using this problem type SHOULD populate both members, with the value of expected-offset taken from the upload resource and the value of provided-offset taken from the upload append request. The following example shows an example response, where the resource's offset was 100, but the client attempted to append at offset 200:

Completed Upload This section defines the "https://iana.org/assignments/http-problem-types#completed-upload" problem type . A server can use this problem type when responding to an upload append request () to indicate that the upload has already been completed and cannot be modified. The following example shows an example response:

Inconsistent Length This section defines the "https://iana.org/assignments/http-problem-types#inconsistent-upload-length" problem type . A server can use this problem type when responding to an upload creation () or upload append request () to indicate that the request includes inconsistent upload length values, as described in . The following example shows an example response:

Content Codings Since the codings listed in Content-Encoding are a characteristic of the representation (see ), both the client and the server always compute the values for Upload-Offset and optionally Upload-Length on the content coded data (that is, the representation data). Moreover, the content codings are retained throughout the entire upload, meaning that the server is not required to decode the representation data to support resumable uploads. See for more information.

Transfer Codings Unlike Content-Encoding (see ), Transfer-Encoding (see ) is a property of the message, not of the representation. Moreover, transfer codings can be applied in transit (e.g., by proxies). This means that a client does not have to consider the transfer codings to compute the upload offset, while a server is responsible for transfer decoding the message before computing the upload offset. The same applies to the value of Upload-Length. Please note that the Content-Length header field cannot be used in conjunction with the Transfer-Encoding header field.

Integrity Digests The integrity of an entire upload or individual upload requests can be verifying using digests from .

Representation Digests Representation digests help verify the integrity of the entire representation data that has been uploaded so far, which might strech across multiple requests. If the client knows the integrity digest of the entire representation data before creating an upload resource, it can include the Repr-Digest header field when creating an upload (). Once the upload is completed, the server can compute the integrity digest of the received representation data and compare it to the provided digest. If the digests don't match, the server SHOULD consider the upload failed, not process the representation further, and signal the failure to the client. This way, the integrity of the entire representation data can be protected. Alternatively, when creating an upload (), the client can ask the server to compute and return the integrity digests using a Want-Repr-Digest field conveying the preferred algorithms. The response SHOULD include at least one of the requested digests, but might not include it. The server SHOULD compute the representation digests using the preferred algorithms once the upload is complete and include the corresponding Repr-Digest header field in the response. Alternatively, the server can compute the digest continuously during the upload and include the Repr-Digest header field in responses to upload creation () and upload appending requests () even when the upload is not completed yet. This allows the client to simultaneously compute the digest of the transmitted representation data, compare its digest to the server's digest, and spot data integrity issues. If an upload is spread across multiple requests, data integrity issues can be found even before the upload is fully completed.

Content Digests Content digests help verify the integrity of the content in an individual request. If the client knows the integrity digest of the content from an upload creation () or upload appending () request, it can include the Content-Digest header field in the request. Once the content has been received, the server can compute the integrity digest of the received content and compare it to the provided digest. If the digests don't match the server SHOULD consider the transfer failed, not append the content to the upload resource, and signal the failure to the client. This way, the integrity of an individual request can be protected.

Responses to Uploads HTTP uploads often not only transfer a representation to the server but also send back information to the client. For resumable uploads, this works similar to conventional HTTP uploads. The server can include information in the response to the request, which transferred the remaining representation data and included the Upload-Complete: ?1 header field. Clients can regard this response as the final response for the entire upload, as detailed in and . However, due to network interruptions, the upload resource might receive the remaining representation data, complete the upload, and send a response to the client, which then does not receive the response. The client can learn that the upload is complete by retrieving its state (), but resumable uploads as defined in this document do not offer functionality to recover the missed response. To address this issue, the server can send the desired information in header fields, which are included in both the final response and responses when the client fetches the upload state via a HEAD request (). This way, the client can retrieve the information from the header even if it failed to receive the final response. If the piece of information is too large for header fields, the server could make it available as a separate resource for retrieval and point the client to its URI in an appropriate header field.

Upload Strategies The definition of the upload creation request () provides the client with flexibility to choose whether the representation data is fully or partially transferred in the first request, or if no representation data is included at all. Which behavior is best largely depends on the client's capabilities, its intention to avoid data re-transmission, and its knowledge about the server's support for resumable uploads. The following subsections describe two typical upload strategies that are suited for common environments. Note that these modes are never explicitly communicated to the server and clients are not required to stick to one strategy, but can mix and adapt them to their needs.

Optimistic Upload Creation An "optimistic upload creation" can be used independent of the client's knowledge about the server's support for resumable uploads. However, the client must be capable of handling and processing interim responses. An upload creation request then includes the full representation data because the client anticipates that it will be transferred without interruptions or resumed if an interruption occurs. The benefit of this method is that if the upload creation request succeeded, the representation data was transferred in a single request without additional round trips. A possible drawback is that the client might be unable to resume an upload. If an upload is interrupted before the client received a 104 (Upload Resumption Supported) interim response with the upload resource's URI, the client cannot resume that upload due to the missing URI. The interim response might not be received if the interruption happens too early in the message exchange, the server does not support resumable uploads at all, the server does not support sending the 104 (Upload Resumption Supported) interim response, or an intermediary dropped the interim response. Without a 104 response, the client needs to either treat the upload as failed or retry the entire upload creation request if this is allowed by the application. A client might wait for a limited duration to receive a 104 (Upload Resumption Supported) interim response before starting to transmit the request content. This way, the client can learn about the resource's support for resumable uploads and/or the upload resource's URI. This is conceptually similar to how a client might wait for a 100 (Continue) interim response (see ) before committing to work.

Upgrading To Resumable Uploads Optimistic upload creation allows clients and servers to automatically upgrade non-resumable uploads to resumable ones. In a non-resumable upload, the representation is transferred in a single request, usually POST or PUT, without any ability to resume from interruptions. The client can offer the server to upgrade such a request to a resumable upload by adding the Upload-Complete: ?1 header field to the original request. The Upload-Length header field SHOULD be added if the representation data's length is known upfront. The request is not changed otherwise. A server that supports resumable uploads at the target URI can create an upload resource and send its URI in a 104 (Upload Resumption Supported) interim response for the client to resume the upload after interruptions. A server that does not support resumable uploads or does not want to upgrade to a resumable upload for this request ignores the Upload-Complete: ?1 header. The transfer then falls back to a non-resumable upload without additional cost. This upgrade can also be performed transparently by a library or program that acts as a HTTP client by sending requests on behalf of a user. When the user instructs the client to send a non-resumable request, the client can perform the upgrade transparently and handle potential interruptions and resumptions under the hood without involving the user. The last response received by the client is considered the response for the entire upload and should be provided to the user.

Careful Upload Creation For a "careful upload creation" the client knows that the server supports resumable uploads and sends an empty upload creation request without including any representation data. Upon successful response reception, the client can use the included upload resource URI to transmit the representation data () and resume the upload at any stage if an interruption occurs. The client should inspect the response for the Upload-Limit header field, which would indicate limits applying to the remaining upload procedure. The retransmission of representation data or the ultimate upload failure that can happen with an "optimistic upload creation" is therefore avoided at the expense of an additional request that does not carry representation data. This approach is best suited if the client cannot receive interim responses, e.g. due to a limitation in the provided HTTP interface, or if large representations are transferred where the cost of the additional request is minuscule compared to the effort of transferring the representation itself.

Security Considerations The upload resource URI is the identifier used for modifying the upload. Without further protection of this URI, an attacker may obtain information about an upload, append data to it, or cancel it. To prevent this, the server SHOULD ensure that only authorized clients can access the upload resource. To reduce the risk of unauthorized access, it is RECOMMENDED to generate upload resource URIs in such a way that makes it hard to be guessed by unauthorized clients. In addition, servers may embed information about the storage or processing location of the uploaded representation in the upload resource URI to make routing requests more efficient. If so, they MUST ensure that no internal information is leaked in the URI that is not intended to be exposed. Uploaded representation data and its metadata are untrusted input. Server operators have to be careful of where uploaded data is written and subsequently accessed, especially if the operations cause the representation to be processed or executed by the server. In addition, metadata MUST be validated and/or sanitized if the server takes its values into consideration for processing or storing the representation. Some servers or intermediaries provide scanning of content uploaded by clients. Any scanning mechanism that relies on receiving a complete representation in a single request message can be defeated by resumable uploads because content can be split across multiple messages. Servers or intermediaries wishing to perform content scanning SHOULD consider how resumable uploads can circumvent scanning and take appropriate measures. Possible strategies include waiting for the upload to complete before scanning the entire representation, or disabling resumable uploads. Resumable uploads are vulnerable to Slowloris-style attacks . A malicious client may create upload resources and keep them alive by regularly sending PATCH requests with no or small content to the upload resources. This could be abused to exhaust server resources by creating and holding open uploads indefinitely with minimal work. Servers SHOULD provide mitigations for Slowloris attacks, such as increasing the maximum number of clients the server will allow, limiting the number of uploads a single client is allowed to make, imposing restrictions on the minimum transfer speed an upload is allowed to have, and restricting the length of time an upload resource can exist.

IANA Considerations

HTTP Fields IANA is asked to register the following entries in the "Hypertext Transfer Protocol (HTTP) Field Name Registry":

Field Name	Status	Structured Type	Reference
Upload-Offset	permanent	Item	of this document
Upload-Complete	permanent	Item	of this document
Upload-Length	permanent	Item	of this document
Upload-Limit	permanent	Dictionary	of this document

HTTP Status Code IANA is asked to register the following entry in the "HTTP Status Codes" registry:

Value:

104 (suggested value)

Description:

Upload Resumption Supported

Specification:

of this document

Media Type IANA is asked to register the following entry in the "Media Types" registry:

Type name:

application

Subtype name:

partial-upload

Required parameters:

N/A

Optional parameters:

N/A

Encoding considerations:

binary

Security considerations:

see of this document

Interoperability considerations:

N/A

Published specification:

of this document

Applications that use this media type:

Applications that transfer files over unreliable networks or want pause- and resumable uploads.

Fragment identifier considerations:

N/A

Additional information:

• Deprecated alias names for this type: N/A
• Magic number(s): N/A
• File extension(s): N/A
• Macintosh file type code(s): N/A
• Windows Clipboard Name: N/A

Person and email address to contact for further information:

See the Authors' Addresses section of this document.

Intended usage:

COMMON

Restrictions on usage:

N/A

Author:

See the Authors' Addresses section of this document.

Change controller:

IETF

HTTP Problem Types IANA is asked to register the following entry in the "HTTP Problem Types" registry:

Type URI:

https://iana.org/assignments/http-problem-types#mismatching-upload-offset

Title:

Mismatching Upload Offset

Recommended HTTP status code:

409

Reference:

of this document

IANA is asked to register the following entry in the "HTTP Problem Types" registry:

Type URI:

https://iana.org/assignments/http-problem-types#completed-upload

Title:

Upload Is Completed

Recommended HTTP status code:

400

Reference:

of this document

IANA is asked to register the following entry in the "HTTP Problem Types" registry:

Type URI:

https://iana.org/assignments/http-problem-types#inconsistent-upload-length

Title:

Inconsistent Upload Length Values

Recommended HTTP status code:

400

Reference:

of this document

References Normative References The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document describes the overall architecture of HTTP, establishes common terminology, and defines aspects of the protocol that are shared by all versions. In this definition are core protocol elements, extensibility mechanisms, and the "http" and "https" Uniform Resource Identifier (URI) schemes. This document updates RFC 3864 and obsoletes RFCs 2818, 7231, 7232, 7233, 7235, 7538, 7615, 7694, and portions of 7230. The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document defines HTTP caches and the associated header fields that control cache behavior or indicate cacheable response messages. This document obsoletes RFC 7234. The Hypertext Transfer Protocol (HTTP) is a stateless application-level protocol for distributed, collaborative, hypertext information systems. This document specifies the HTTP/1.1 message syntax, message parsing, connection management, and related security concerns. This document obsoletes portions of RFC 7230. This document describes a set of data types and associated algorithms that are intended to make it easier and safer to define and handle HTTP header and trailer fields, known as "Structured Fields", "Structured Headers", or "Structured Trailers". It is intended for use by specifications of new HTTP fields. This document obsoletes RFC 8941. Several applications extending the Hypertext Transfer Protocol (HTTP) require a feature to do partial resource modification. The existing HTTP PUT method only allows a complete replacement of a document. This proposal adds a new HTTP method, PATCH, to modify an existing HTTP resource. [STANDARDS-TRACK] This document defines a "problem detail" to carry machine-readable details of errors in HTTP response content to avoid the need to define new error response formats for HTTP APIs. This document obsoletes RFC 7807. This document defines HTTP fields that support integrity digests. The Content-Digest field can be used for the integrity of HTTP message content. The Repr-Digest field can be used for the integrity of HTTP representations. Want-Content-Digest and Want-Repr-Digest can be used to indicate a sender's interest and preferences for receiving the respective Integrity fields. This document obsoletes RFC 3230 and the Digest and Want-Digest HTTP fields. RFC 2616 defines the Content-Disposition response header field, but points out that it is not part of the HTTP/1.1 Standard. This specification takes over the definition and registration of Content-Disposition, as used in HTTP, and clarifies internationalization aspects. [STANDARDS-TRACK] In many standards track documents several words are used to signify the requirements in the specification. These words are often capitalized. This document defines these words as they should be interpreted in IETF documents. This document specifies an Internet Best Current Practices for the Internet Community, and requests discussion and suggestions for improvements. RFC 2119 specifies common key words that may be used in protocol specifications. This document aims to reduce the ambiguity by clarifying that only UPPERCASE usage of the key words have the defined special meanings. Informative References This document defines two strategies for handling long lines in width-bounded text content. One strategy, called the "single backslash" strategy, is based on the historical use of a single backslash ('\') character to indicate where line-folding has occurred, with the continuation occurring with the first character that is not a space character (' ') on the next line. The second strategy, called the "double backslash" strategy, extends the first strategy by adding a second backslash character to identify where the continuation begins and is thereby able to handle cases not supported by the first strategy. Both strategies use a self-describing header enabling automated reconstitution of the original content.

Changes

Since draft-ietf-httpbis-resumable-upload-08

• Clarify definitions of new header fields.
• Make handling of OPTIONS * optional.
• Require server to announce limits using Upload-Limit.
• Require clients to adhere to known limits.
• Rephrase requirements for concurrency handling, focusing on the outcome.
• Remove requirement for 204 status code for DELETE responses.
• Increase the draft interop version.
• Add section about 104 status code.
• Rephrase recommendation for sending information back to client.

Since draft-ietf-httpbis-resumable-upload-07

• Clarify server handling when upload length is exceeded.
• Extend security considerations about upload resource URIs, representation metadata, and untrusted inputs.
• Allow clients to retry for appropriate 4xx responses.

Since draft-ietf-httpbis-resumable-upload-06

• Minor editorial improvements to introduction and examples.
• Define structured types for new header fields.

Since draft-ietf-httpbis-resumable-upload-05

• Increase the draft interop version.
• Numerous editorial changes.
• Rename expires limit to max-age.
• Require Upload-Complete, but not Upload-Offset or Upload-Limit, for append responses.
• Add problem type for inconsistent length values.
• Reduce use of "file" in favor of "representation".

Since draft-ietf-httpbis-resumable-upload-04

• Clarify implications of Upload-Limit header.
• Allow client to fetch upload limits upfront via OPTIONS.
• Add guidance on upload creation strategy.
• Add Upload-Length header to indicate length during creation.
• Describe possible usage of Want-Repr-Digest.

Since draft-ietf-httpbis-resumable-upload-03

• Add note about Content-Location for referring to subsequent resources.
• Require application/partial-upload for appending to uploads.
• Explain handling of content and transfer codings.
• Add problem types for mismatching offsets and completed uploads.
• Clarify that completed uploads must not be appended to.
• Describe interaction with Digest Fields from RFC9530.
• Require that upload offset does not decrease over time.
• Add Upload-Limit header field.
• Increase the draft interop version.

Since draft-ietf-httpbis-resumable-upload-02

• Add upload progress notifications via informational responses.
• Add security consideration regarding request filtering.
• Explain the use of empty requests for creation uploads and appending.
• Extend security consideration to include resource exhaustion attacks.
• Allow 200 status codes for offset retrieval.
• Increase the draft interop version.

Since draft-ietf-httpbis-resumable-upload-01

• Replace Upload-Incomplete header with Upload-Complete.
• Replace terminology about procedures with HTTP resources.
• Increase the draft interop version.

Since draft-ietf-httpbis-resumable-upload-00

• Remove Upload-Token and instead use Server-generated upload URL for upload identification.
• Require the Upload-Incomplete header field in Upload Creation Procedure.
• Increase the draft interop version.

Since draft-tus-httpbis-resumable-uploads-protocol-02 None

Since draft-tus-httpbis-resumable-uploads-protocol-01

• Clarifying backtracking and preventing skipping ahead during the Offset Receiving Procedure.
• Clients auto-retry 404 is no longer allowed.

Since draft-tus-httpbis-resumable-uploads-protocol-00

• Split the Upload Transfer Procedure into the Upload Creation Procedure and the Upload Appending Procedure.

Draft Version Identification To assist the development of implementations and interoperability testing while this document is still a draft, an interop version is defined. Implementations of this draft use the interop version to identify the iteration of the draft that they implement. The interop version is bumped for breaking changes. The current interop version is 8. Client implementations of draft versions of the protocol MUST send a header field Upload-Draft-Interop-Version with the interop version as its value to its requests. The Upload-Draft-Interop-Version field value is an Integer. Server implementations of draft versions of the protocol MUST NOT send a 104 (Upload Resumption Supported) informational response when the interop version indicated by the Upload-Draft-Interop-Version header field in the request is missing or mismatching. Server implementations of draft versions of the protocol MUST also send a header field Upload-Draft-Interop-Version with the interop version as its value to the 104 (Upload Resumption Supported) informational response. Client implementations of draft versions of the protocol MUST ignore a 104 (Upload Resumption Supported) informational response with missing or mismatching interop version indicated by the Upload-Draft-Interop-Version header field. The reason both the client and the server are sending and checking the draft version is to ensure that implementations of the final RFC will not accidentally interop with draft implementations, as they will not check the existence of the Upload-Draft-Interop-Version header field.

Acknowledgments This document is based on an Internet-Draft specification written by , , and the authors of this document. The tus v1 protocol is a specification for a resumable file upload protocol over HTTP. It inspired the early design of this protocol. Members of the tus community helped significantly in the process of bringing this work to the IETF. The authors would like to thank for substantive contributions to the text, and and for their thorough reviews of the document.