 Sajin
|
46092807eb
epoch_duration removed from enclave_api.h (clients don't need epoch_durations). Epoch_duration taken as a command line parameter for experiment script to tune that depending on the experiment. Log connected clients to confirm the number of active clients
|
il y a 1 an |
|
Sajin
|
9775d8e843
Fix for initializing NodeIO variables to 0. (Resolves the server crash when instantiating experiments with >18 servers)
|
il y a 1 an |
|
Sajin Sasy
|
3753edc35a
Fixed Bad Address error. (send_client_mailbox was using msgbundle_size instead of mailbox_size)
|
il y a 1 an |
|
Sajin Sasy
|
f15b914acd
Minor tweaks for nomenclature consistency of msgbundles and mailboxes
|
il y a 1 an |
|
Sajin Sasy
|
5a46a77779
Fixing and seperating sizes on msgbundles (the messages clients send each epoch, which has tokens alongside the messages) and mailbox (the messages clients receive each epoch, which do NOT have tokens)
|
il y a 1 an |
|
Sajin Sasy
|
7f14b5308a
Storage servers send back token bundles and messages every epoch to clients that have authenticated
|
il y a 1 an |
|
Sajin Sasy
|
0f2f054510
Storage servers provide first epoch token bundles to clients when they authenticate. (Currently without using epoch no in token computation)
|
il y a 1 an |
|
Sajin Sasy
|
a6eb344280
If authentication fails, don't accept client messages on that socket
|
il y a 1 an |
|
Sajin Sasy
|
c762dc289c
Clients authenticate with epoch number as IV. Client messages (both authenticate and message_bundles) no longer use headers
|
il y a 1 an |
|
Sajin Sasy
|
4e6f8ee07d
Ingestion processing uses the ingbuf in route_state now. Removed unnecessary memcpy's in ingestion processing. Clients maintain a SimID and a ClientID, and send messages to the correct ClientID format.
|
il y a 1 an |
|
Sajin Sasy
|
1aeb41385b
Ingestion servers handle client connections asynchronously correctly (using async_read now). Clients send CLIENTS_AUTHENTICATE message first, then several CLIENT_MESSAGE_BUNDLE messages
|
il y a 1 an |
|
Sajin Sasy
|
461aca4bf9
Revert "Syncing to debug acceptor problem"
|
il y a 1 an |
|
Sajin Sasy
|
192acaccb0
Syncing to debug acceptor problem
|
il y a 1 an |
|
Sajin Sasy
|
1f36f457c3
Merged main
|
il y a 1 an |
|
Sajin Sasy
|
bda0918bd3
Ingestion processes msgbundle, and queues messages into a msgBuffer
|
il y a 1 an |
|
Sajin Sasy
|
aa86ecf9c1
Client -> Ingestion send (unencrypted) message bundles. Ingestion servers queue them into a msgBuffer
|
il y a 1 an |
 Ian Goldberg
|
61fe90b6d6
Start round 2
|
il y a 1 an |
|
Sajin Sasy
|
9422c28caf
Client single async connect to Server (App)
|
il y a 1 an |
|
Ian Goldberg
|
518cdb8a14
The client acceptor socket needs to be long lived
|
il y a 1 an |
|
Sajin Sasy
|
7fe40246f7
Sync state to debug client->server async connection
|
il y a 1 an |
|
Ian Goldberg
|
68c67d8621
More towards private routing
|
il y a 1 an |
|
Ian Goldberg
|
307b4f7823
Client sockets had the wrong node number in the NodeIO struct
|
il y a 1 an |
|
Ian Goldberg
|
9e22973195
The enclave side of the untrusted app to enclave I/O
|
il y a 1 an |
|
Ian Goldberg
|
f98d4bd65c
Change MAXCHUNKSIZE to FRAME_SIZE and move it to enclave_api.h
|
il y a 1 an |
|
Ian Goldberg
|
6b3139b84c
Forward received data into the enclave
|
il y a 1 an |
|
Ian Goldberg
|
dd2f2be3d0
Actually send outgoing data from the enclave
|
il y a 1 an |
|
Ian Goldberg
|
ae76624b90
Stubs for the I/O functions in and out of the enclave
|
il y a 1 an |
|
Ian Goldberg
|
d17365c8ea
Rename App/config.* to App/appconfig.*
|
il y a 1 an |
|
Ian Goldberg
|
12a9d0aeb8
Use defined constants for the command type bytes rather than hardcoding 0x01, etc.
|
il y a 1 an |
|
Ian Goldberg
|
3a506cf620
Nodes can now asynchronously receive data using callbacks
|
il y a 1 an |
