In 603 the Roman Emperor Phokas restored Smaragdus to his former position as Exarch of Ravenna. Smaragdus inherited a war with the Lombards from his predecessor Callinicus, Agilulf entered Mantua on 13 September, and Vulturina surrendered about the same time, leaving much of northern Italy in Lombard hands. Smaragdus was forced to release his hostages in April 605 in order to gain peace. The peace with the Lombards held for the rest of his administration till he was replaced in 611 shortly after his benefactor's assassination by Heraclius...
WI Smaragdus rose in rebelion against Heraclius in 610 to revenge the murder of Phokas? Could the Empire handled another civil war with the Persians fooling around the Eastern borders?
How is that altering History? Any thoughts?
This means pretty much the end to Byzantine Italy, if you ask me. Until at least 625 or so, presuming similar outcome of the war in the East, Byzantium is in no position to reinforce its authority in Italy. And since the Arab invasion is almost a certainty by then, given that the POD is not likely to change Muhammad's eventual fate... there are some interesting issues down the line.
In OTL, during Constans II's disastrous defeats against the Arabs, he moved his court to Syracuse, and seriously considered moving to Italy in full force. Here, Roman authority in Italy is likely to be much less, and while I doubt the revolt of the Exarch is going to reach as far down as Sicily, it is likely to be seen as a far less safe place than OTL. Therefore, Constans II or his equivalent is unlikely to consider Sicily as a refuge of the Roman power.
Of course, my assumption is that chaos theory does not result in some unpredictable butterflies - i.e. Arabs being less successful, stray arrow hitting Heraclius and ending the dynasty before it began, Romans being more successful against the Persians or the other way around, etc etc. I am assuming that the large-scale movements in history occur very similar to OTL in order to arrive at the above opinion.