Don't really know who is broadcasting exactly, but it's doing so daily. And I don't know how can I stop a transaction. I wanted to send btc from my blockchain wallet to btce , so who should I ask to stop the broadcasting?
That is the strange part. AFAIK Blockchain has no method for manually killing a tx but they normally do stop broadcasting it after a few days if it doesn't confirm.
The other strange thing is how/why Blockchain allowed you to create that tx outside of some expert raw transaction mode as it had almost no chance of being confirmed in any timely manner. Low priority, fee below the requirement to relay, and dust output. You hit the trifecta of "probably not going to make it into the next block". A good wallet should prevent users from making these types of questionable tx (unless they explicitly request it). The Bitcoin core client for example would never have made this tx.
You had 7.91776 mBTC (0.00791776 BTC) in inputs.
The tx created sent 7.9 mBTC, which left 0.01776 mBTC remaining.
It paid 0.01 mBTC as a fee (which was pointless at the time, v0.9 will change that).
That left 0.00776 mBTC which is below the dust threshold of 0.05430 mBTC.
No possible way to make that tx have a worse chance of confirming. It isn't your fault (unless you used some manual mode) blockchain.info should never have created that. Do you remember exactly how you made this tx in blockchain.info?
Instead the QT client would have showed you that you have the same 7.91776 mBTC in the wallet (I will assume there are no other outputs available). It knows the tx will be low priority so it will force a fee on 0.1 mBTC (to avoid this scenario). Thus it wouldn't let you send more than 7.81776 mBTC. That tx would have been confirmed within a few blocks max. If you wanted to send slightly less (say 7.8 mBTC) that would have left 0.01776 after the fee as your change output which would violate the dust rule so the wallet would instead not create that output and increase the fee to the miner to avoid making the dust (i.e 0.11776 mBTC to miner instead of default 0.1 mBTC).
If blockchain.info had paid insufficient fee but not created dust it probably would be confirmed by now (at least one miner must be running v0.9). What I mean by that is instead of your outputs being:
7.90000 mBTC - what you were trying to spend/transfer
0.00776 mBTC - the change
0.01000 mBTC - the fee to miner
it would instead be
7.90000 mBTC - what you were trying to spend/transfer
0.01776 mBTC - the fee to miner
and no change output
that tx would have violated the relay rules (and min miner fee rules) for v0.8x but those are being reduced to 0.01 mBTC in v0.9 so you would have "lucked out" due to the rule change. However in your case that dust output is the killer. It makes the tx non-standard and unless a miner uses some custom logic to pick non-standard txs it will never be part of the next block. You may want to try submitting the tx to eligus pool directly. They support non-standard txs but I am not sure they will take it with that fee.
The tx is low priority so it should have included the min fee of 0.1 mBTC that would leave 0.00781776 BTC