Meetings2011-06-27 » History » Version 3
J. Moringen, 06/27/2011 11:31 AM
1 | 1 | J. Moringen | h1. Meetings2011-06-27 |
---|---|---|---|
2 | 1 | J. Moringen | |
3 | 1 | J. Moringen | h2. Cross-Compilation for an ARM-core |
4 | 1 | J. Moringen | |
5 | 1 | J. Moringen | * OMAP3 processor (made by TI) |
6 | 1 | J. Moringen | |
7 | 1 | J. Moringen | h2. Scenario |
8 | 1 | J. Moringen | |
9 | 1 | J. Moringen | * Mobile Robot |
10 | 1 | J. Moringen | * CAN-Bus -> Ethernet gateway |
11 | 1 | J. Moringen | * CAN-Messages |
12 | 1 | J. Moringen | ** ID + 8 byte payload |
13 | 1 | J. Moringen | |
14 | 1 | J. Moringen | * Virtual CAN-Bus |
15 | 1 | J. Moringen | ** Socket-interface to simulated CAN-Bus (in Linux) |
16 | 1 | J. Moringen | ** For simulation purposes (until read hardware becomes available) |
17 | 1 | J. Moringen | |
18 | 1 | J. Moringen | * Use-cases |
19 | 1 | J. Moringen | ** Transmitting infrared sensor readouts "out of" robots |
20 | 2 | J. Moringen | ** Pub/Sub on CAN-Bus level |
21 | 2 | J. Moringen | |
22 | 2 | J. Moringen | h2. RSB Discussion |
23 | 2 | J. Moringen | |
24 | 2 | J. Moringen | * Mapping of RSB channel designators to Spread groups |
25 | 2 | J. Moringen | ** MD5-hash |
26 | 2 | J. Moringen | |
27 | 2 | J. Moringen | * TODO visualization of channel hierarchy |
28 | 2 | J. Moringen | |
29 | 2 | J. Moringen | h2. Avoiding Sending of Events when there is no Receiver |
30 | 2 | J. Moringen | |
31 | 2 | J. Moringen | * The relevant use case is disabling and enabling computations based on presence of consumers |
32 | 2 | J. Moringen | * A callback interface for the "receiver appeared/disappeared" events would be sufficient |
33 | 2 | J. Moringen | |
34 | 2 | J. Moringen | * Event-based notification about receivers (no additional callbacks) |
35 | 2 | J. Moringen | ** Probably in the Participant base-class |
36 | 2 | J. Moringen | ** Handler-infrastructure should be used to discriminate data vs. meta-data events |
37 | 2 | J. Moringen | |
38 | 2 | J. Moringen | * Logging and introspection could be a problem for this (since a logger on "/" may trigger lots of stuff) |
39 | 2 | J. Moringen | |
40 | 2 | J. Moringen | h2. Deployment |
41 | 2 | J. Moringen | |
42 | 2 | J. Moringen | * Node configuration changes dynamically |
43 | 2 | J. Moringen | * Is Spread's configuration mechanism a problem? probably not |
44 | 2 | J. Moringen | * Scenario could scale up to 64 nodes |
45 | 2 | J. Moringen | * Plug and play would be useful in some use cases |
46 | 2 | J. Moringen | ** E.g. non-expert users |
47 | 2 | J. Moringen | ** But Spread may be a bit of a problem here |
48 | 2 | J. Moringen | ** Initial approach: find a working network configuration that works well with Spread |
49 | 2 | J. Moringen | ** Later: maybe without Spread |