4 A rough list of stuff that is going to be done to SILC after 1.0 or at
7 o Implement the defined SilcDH API. The definition is in
8 lib/silccrypt/silcdh.h.
10 o X.509 certificate support. SILC protocol supports certificates and
11 it would be great to have support for them. This is a big task as
12 support has to be made for ASN.1 as well. I've looked into OpenSSL
13 package as it has X.509 certificate support (and ASN.1 as well).
14 The code does not look very good to my eye but it has some potentials.
15 This should be looked at more closely.
17 Naturally own SILC Certificate API has to be defined regardles what
18 the actual X.509 library is (OpenSSL X.509 or something else). Other
19 choice is to write own X.509 library but I'm not going to do it -
20 I can help to migrate the OpenSSL X.509 into SILC and I can help if
21 someone would like to write the X.509 library - but I'm not going
22 to start writing one myself. Anyhow, the OpenSSL X.509 lib should
25 Other package that should be checked is the NSS's X509 library,
26 which I like more over OpenSSL package.
28 o SSH2 public keys support, allowing the use of SSH2 public keys in
31 o OpenPGP certificate support, allowing the use of PGP public keys
34 o Compression routines are missing. The protocol supports packet
35 compression thus it must be implemented. SILC Zip API must be
38 o Optimizations in Libraries
40 o There is currently three (3) allocations per packet in the
41 silc_packet_receive_process, which is used to process and
42 dispatch all packets in the packet queue to the parser callback
43 function. First allocation is for parse_ctx, second for the
44 SilcPacketContext, and third for packet->buffer where the actual
47 The parse_ctx allocation can be removed by adding it as a
48 structure to the SilcPacketContext. When the SilcPacketContext
49 is allocated there is space for the parse context already.
51 The silc_packet_context_alloc could have a free list of
52 packet contexts. If free packet context is found from the list
53 it is returned instead of allocating a new one. The library
54 could at first allocate them and save them to the free list
55 until enough contexts for smooth processing exists in the list.
56 This would remove a big allocation since the structure is
57 quite big, and even bigger if it would include the parse_ctx.
59 The packet->buffer can be optimized too if the SilcBuffer
60 interface would support free lists as well. Maybe such could
61 be done in the same way as for SilcPacketContext. The
62 silc_buffer_alloc would check free list before actually
63 allocating new memory. Since the packets in the SILC protocol
64 usually are about the same size (due to padding) it would be
65 easy to find suitable size buffer from the free list very
68 These naturally cause the overal memory consumption to grow
69 but would take away many allocations that can be done several
72 o Move the actual file descriptor task callback (the callback that
73 handles the incoming data, outgoing data etc, that is implemnted
74 in server and client separately (silc_server_packet_process and
75 silc_client_packet_proces)) to the low level socket connection
76 handling routines, and create an interface where the application
77 can register a callbacks for incoming data, outoing data and EOF
78 receiving, which the library will call when necessary. This way
79 we can move the data handling in one place.
81 o Add silc_id_str2id to accept the destination buffer as argument
82 and thus not require any memory allocation. Same will happen
83 with silc_id_payload_* functions.
85 o Remove the `truelen' field from SilcBuffer as it is entirely
86 redundant since we can get the true length of the buffer by
87 doing buffer->end - buffer->header. Add SILC_BUFFER_TRUELEN
88 macro instead. Consider also removing `len' field too since
89 it effectively is buffer->tail - buffer->data, and adding
90 SILC_BUFFER_LEN macro can do the same. These would save
91 totally 8 bytes of memory per buffer.
93 o Scheduler can be optimized for FD tasks by changing the fd_queue
94 to SilcHashTable instead of using linked list. We need to do
95 one-to-one mapping of FD to task and hash table is more efficient
98 Also redefine the silc_select to perhaps return a separate
99 structure of the events that actually occurred, instead of
100 returning the events in the fd_list which is then traversed
101 in the generic code to find the changed events. This can be
102 made faster by having own struct which includes only the
103 changed events, thus the tarversing is faster since the whole
104 fd_list is not traversed anymore (it is still traversed in the
105 silc_select but at least it removes one extra tarversing later
108 o Optimizations in Server
110 o Remove the big switch statement from the function
111 silc_server_packet_parse_type and replace it with predefined
112 table of function pointers where each of the slot in table
113 represents the packet type value.
115 Same could be done with notify packets which has big switch
116 statement too. Same kind of table of notify callbacks could be
119 o The parser callback in the server will add a timeout task for
120 all packets. It will require registering and allocating a
121 new task to the SilcSchedule. Maybe, at least, for server
122 and router packets the parser would be called immediately
123 instead of adding it to the scheduler with 0 timeout. It
124 should be analyzed too how slow the task registering process
125 actually is, and find out ways to optimize it.
127 o The SERVER_SIGNOFF notify handing is not optimal, because it'll
128 cause sending of multiple SIGNOFF notify's instead of the one
129 SERVER_SIGNOFF notify that the server received. This should be
130 optimized so that the only SERVER_SIGNOFF is sent and not
131 SIGNOFF of notify at all (using SIGNOFF takes the idea about
132 SERVER_SIGNOFF away entirely).
134 o Add SilcAsyncOperation to utility library. Any function that takes
135 callback as an argument must return SilcAsyncOperation.
139 o Cipher optimizations (asm, that this) at least for i386 would be nice.
141 o Add builtin SOCKS and HTTP Proxy support, well the SOCKS at least.
142 SILC currently supports SOCKS4 and SOCKS5 but it needs to be compiled