You mean pool connection or my connection?

I will try restart router and PC, have you some other ideas?

that is the last option

Lot's of pools under DDoS rolling attack it seems.

 It depends where you are looking to see your hashrate. If it is the estimate from the account page - that is affected by the last 10 rounds, for all workers together, or based on shares in current round for each worker. If you join a round already in progress, that will be a poor estimate. If your hashrate on your actual device as seen from miner output, well now that would be weird!

It's a scary new day:

http://arstechnica.com/security/2013/04/fueled-by-super-botnets-ddos-attacks-grow-meaner-and-ever-more-powerful/?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+arstechnica%2Findex+%28Ars+Technica+-+All+content%29

I know it's slow. :-)
Yes, i did that, I'm mining on backup pool.

Better pools...

OK, Slush's is still having issues with DDos ( and aftermath of said same ) and I am using GUIMiner to do CPU mining ( don't go there, look at my other posts first if ya gotta )

1) restart PC
2) copy the tabs you are using for the Stratum and set to slush's pool in the drop down from either solo or other ( if that is how you were mining )
3) copy your miner to a similarly named directory and make a ' new cpu miner ' ( no, not actually going to use one, i personaly use 2 and this sometimes helps )
4) find all the settings you need to pass to your miner in the command line ( NOTEPAD is your friend at this point ) and put the full directoy,miner name and those settings in at the top for where you would use a CPU miner and uncheck one of the CPU afinity boxes.
If you got quad, you will need to do this 2 more times and 2 more directories so you end up with one miner per CPU affinity
5) SAVE SETTINGS
6) run them and use the summary tab ( use the X at the top right to kill the tabs you do not need/want IF separating the CPU affinity did not help )
Look at the bottom right, if it's over 1.1Mhash/sec you are doing better than me. 

Eventualy you will find the correct core per mining directory to max out your hashes ( total hashes not per pool hashes ).  I do recomend stress testing your system with a CPU miner in it's own work tab AND having GPU miners.  There are a ton of process ticks between the sends/recieves from CUDA mining that may get you a bump of a few hundred khash/sec and it all adds up.

If this helps, I do take donations