Open pjcozzi opened 9 years ago
Is anyone interested in this for the draft 1.0 spec? If so, can you please describe your use case?
Requested by @e-andersson.
Example:
"tiles": {
"level1-id" : "0/0/0.b3dm"
},
// ...
"content": {
"tile": "level1-id",
"byteOffset": 0,
"byteLength": 1024
}
This will be a great feature, but am pushing 1.0 to narrow scope.
I know the link in the description alludes to this already, but given that HTTP/2 support is now standard in all browsers, is this still needed? Does it provide any benefit other than reducing the number of tile requests?
Not sure when HTTP/2 support became available in each browser, but keep in mind that not all users are able to use the latest browser.
@austinEng is evaluating HTTP/2 vs. combing vs. both. Results TBA soon.
Awesome, looking forward to the results and methodology (since HTTP/2 server configuration probably comes into play as well)
Would love to do this, but pushing post 1.0 for scope.
This issue came up in discussion today. Some notes:
gRPC
(or other similar library) might be useful to make a server/client architecture that can bundle tiles together in a single request3DTILES_implicit_tiling
in the future, as the availability buffers will be helpful in locating tiles to request on the server.CC @lilleyse @sanjeetsuhag
gRPC (or other similar library) might be useful to make a server/client architecture that can bundle tiles together in a single request
But HTTP/2 can already do this, right?
See When using replacement refinement, can multiple children be combined into one request?.