( all measurements taken from same building with wired ethernet ).
Fiber has a little better latency (and more consistent) than VDSL (14 vs 12 ms, but jitter is more consistent). A good VDSL line is still better (latency and -consistency-wise) than DOCSIS Cable Internet (latency is very high during upload).
Bandwith is not comparable due to different tariffs, but each maxes out the promised Down/Upstreams.
The key is generated from a live ISO, is never stored on disk, and I don't use a printer, so the private key never needs to pass through another system. The downside is that, whenever I need to use it, I have to type the key manually and avoid making any transcription errors. I know this is inconvenient, but I only use the key occasionally for SSH certificate signing, so I consider the trade-off worthwhile.
I came up with two tricks to make the manual transcription less error-prone:
y and g and q which are easily confused when written by hand)For example, I use the following .na
| -- Recommendations to stabilise your Oracle Database XE Edition (tested on 18 XE) | |
| -- limit background processes so that CJQ0 Job does not eat too much RAM (Default: 4000?) (avoid ORA-04036) | |
| ALTER SYSTEM SET job_queue_processes=123; | |
| -- the Number of Space Background Workers (W00n manamged by SMCO) scales with the number of CPU Cores? On 12 core, when each uses 200 MB, PGA is quickly exhausted (avoid ORA-04036) | |
| ALTER SYSTEM SET "_max_spacebg_slaves"=4 SCOPE=BOTH; |
| set serveroutput on | |
| DECLARE | |
| -- $VER = 1cbc41ebb40e5ef52b7f5958adb718a1f3600fb73d717476c8b747e138b47866 | |
| -- script to consume lots of PGA Memory for testing / PGA bombing for session tests - will eventually trigger ORA-04036 | |
| bytesused CONSTANT INTEGER := 1024 * 1024 * /*MB=*/100*7; | |
| TYPE array_t IS | |
| VARRAY(bytesused) OF CHAR(32767); -- 32K per row ~ 32 rows/MB | |
| array array_t := array_t(); | |
| line VARCHAR(32767); | |
| BEGIN |
--> nearly exactly as advertised, 64 kbit/s bidirectional. (Note: looks like you have a certain burst allowance for the first 10 seconds).
You should set smaller blocks for iperf, because otherwise you won't see anything in between:
~$ iperf3 -c speedtest.wtnet.de -p 5204 -t 86 --bidir -V -l 8K -b 64K -t 13 -O 2
iperf 3.20
Linux ubuntu 7.0.0-10-generic #10-Ubuntu SMP PREEMPT_DYNAMIC
Control connection MSS 1428
So you already know that multicast traffic is slowing down your wifi?!
And you're using your unify APs "offline", that is, they have no planned internet connection?
You must configure Gateway and DNS IPs in your APs that are reachable in the local lan.
If you have the instantclient "installed" and want to enable some diagnostic debug output, in the same folder as your tnsnames.ora
create a file sqlnet.ora
echo %TNS_ADMIN%
c:\oracle\network\admin\
file contents (remove comment to make active):
The debian 13.0 default/gnome live dvd,
Although libqrencode4 is installed, it is missing qrencode by default.
qrencode works nicely and supports creating linked/sequenced QRCodes, i.e. "structured Append".
Klarna Festgeld zu verschiedenen Zeitpunkten
2025-06-18:
In 2025, can you max out a 10 Gbps Link on the local Network? What do you need?
-> fast PCIe Ports with >= x4 Lanes.
I set out to test this.
TL;DR: the 1000Base-T is astonishingly very well optimized, has very small latency, 10G is not going to change much.
The 100M and 1G Tests were performed on the same machine, via loopback cable/fiber and all tests without a router.
It is not so common to be able to assign the same IP Subnet to different Interfaces on the same machine, but go see this excellent guide.