1
Media Streaming To-Do List
2
=========================
4
0047: Make the default port something in the 50,00 -> 60,000 range and add a build config option to set it at build time.
7
0046: An daemon level default backup location.
9
0045: Better error reporting for daemon level backup. [RESOLVED: RN141]
11
0044: Rewrite/reformat PBMS documentation. [RESOLVED: RN161]
13
0043: Fix the problem where it is possible for data that was commited at the time of the backup not to be included in the backup.
15
0042: Fix BLOB indexing bug. [Feature removed: RN190]
17
0041: Fix transaction errors when a table or database is dropped. [RESOLVED: RN156]
19
0040: Make data stored in the pnms_variable table persistent. [RESOLVED: RN141]
21
0039: Enable cloud storage. [RESOLVED: RN141]
23
0038: Add Drizzle support.
25
0036: Provide a patch to for MySQL to make all engines PBMS enabled.
27
0035: Provide a patch for InnoDB to make it PBMS enabled.
29
0034: Implement incremental backup.
31
0033: Implement remote repositories.
33
0032: Implement daemon level replication.
36
0031: Implement PBMS_OPTION_TRANSMITION_TIMEOUT.
38
0030: Implement pbms_keep_alive. This could be done at the socket level. The threads could set a timeout value before beginning a read/write operation. A seperater thread would monitor the workers and interupt them if the timeout expired.
42
0025: Port to Windows.
44
0024: Add a system table for HTTP meta data fields.[RESOLVED: RN54]
46
0023: Implement index support for all PBMS system tables.
48
0022: Add support for BLOB access via the BLOB alias.[RESOLVED: RN53] [Feature removed: RN190]
50
0021: Add transaction support.[RESOLVED: RN64]
53
0020: Add built in functions for use in triggers to enable non Media Streaming enabled engines to use PBMS.[RESOLVED: RN45] [Feature removed: RN59]
55
0019: Register a port number to be used as the default HTTP port number used by the daemon.
57
0018: Enable the HTTP server to handle 1000+ connections efficiently.
59
0017: Make the HTTP server more robust so that it can tolerate HTTP connections that do not have the keep-alive set but do not terminate immediately after receiving a reply.
61
0016: Add a connection timeout so that the HTTP server can close idle HTTP connections.
65
0015: Implement index support for the repository tables.
67
0014: Implement UPDATE/INSERT support for the repository tables.[RESOLVED: RN51, RN52]
69
0013: Implement support for sorting the repository tables.[RESOLVED: RN58]
71
0012: Check limits and formats of given system variable values.
73
0011: Implement built-in functions BLOB_DATA() and/or BLOB_URL().
77
0010: Add content types [RESOLVED: RN10].
79
0009: Add create, access and reference timestamps [RESOLVED: RN9].
81
0008: Only open one file handle per file under UNIX.
83
0007: Implement RENAME TABLE with Blobs in the repository [RESOLVED: RN8].
85
0006: Implement DROP TABLE with Blobs in the repository [RESOLVED: RN7].
87
0005: Implement the MyBS compactor thread [RESOLVED: RN6].
89
0004: Implement the MyBS virtual tables [RESOLVED: RN11].
91
0003: Insert reference into a table. Implement reference and dereference BLOB [RESOLVED: RN5].
93
0002: Implement the file handle pool.
95
0001: Implement the temp BLOB list, including the temp list monitoring thread [RESOLVED: RN4].
99
There is still a lot to be done. The first version of the BLOB Streaming daemon only allows download of selected column. The BLOBs must still be INSERTed into the database normally.
101
The next step is add support of the upload of BLOBs. It will then be possible to store BLOBs of any size in the database.
103
The streaming enabled storage daemon will support this by storing BLOBs "out of record". When SELECTing these BLOBs from the server, a URL will be returned instead of the BLOB data. The URL can then be used GET the BLOB via the BLOB Streaming daemon.
105
One of the top priorities is to introduce a security concept for the BLOB Streaming daemon.
107
The current idea is to introduce a database user with a specific name (e.g. "mybs") who's grant profile reflects the columns that can be accessed via the BLOB Streaming daemon.
109
I am still gathering ideas on this and other issue so input is welcome!
112
PrimeBase Technologies GmbH
113
barry.leslie@primebase.com
114
www.blobstreaming.org