From Empire of the Winds: The Global Role of Asia’s Great Archipelago, by Philip Bowring (I.B. Tauris, 2018), Kindle pp. 159-160:
Nusantarian commerce in the western Indian Ocean did not suddenly vanish, leaving the settlements cut off from their roots. Ships from Java and Sumatra continued to play a role in Indian Ocean trade at least until the thirteenth century, not least in the slave trade. In the mid-tenth century an Arab ship encountered off Mozambique a group of raiders described as ‘Waqwaq’. ‘Waqwaq’ was a vague term used by Arabs to denote peoples from the extreme south or east, hence probably Nusantarian. (Waqwaq was the subject of myths about islands where girls grew on trees.)
East Africa was a source of slaves for hundreds of years, with the Baghdad-based Abbasid empire the main market. The Zanj slaves became so numerous that they became a major factor in the long-lasting anti-Abbasid rebellion which led to the sack of Basra in 871. This horrific event was widely written about in near contemporary Arab literature, including Muhammad el-Tabari’s History of Prophets and Kings and Muhammad al-Biruni’s Chronology of Ancient Nations. Some slaves were even sold in China. A Chinese, Zhu Yu, writing around 1100, recorded that wealthy people in Guangzhou employed what they called ‘devil slaves’ from Africa: …
Arab sources from the twelfth century refer to raids on the African coast by boats coming from Madagascar, though it is not clear if these were from settlements in Madagascar or by Nusantarians basing themselves there for the slave trade or piracy. Madagascar was too underpopulated to be a major source of slaves but it may well have been a transit centre and even destination, and was mentioned by a thirteenth century Arab chronicler as the origin for slave shipments to Jeddah and Yemen. (The Dutch had a post at Moroantsetra on the north-east coast from 1641 to 1646 but abandoned it because the local ruler had too few slaves to sell. However, Madagascar was a source of several thousand slaves for the American market till the early nineteenth century.)
The Arabs recorded that the people of Madagascar did not have ocean-going boats but ‘people from Java come in large ships and export products from Madagascar because they speak the same language.’ Those products probably included rice, which was exported to Africa. Nusantarian ships are also known to have visited Sofala, on the coast of Mozambique, the port which gave access to the gold mines of Zimbabwe.
Indian Ocean trade became increasingly dominated by Muslims of various origins – not excluding those from Sumatra. Yet despite the presence for centuries of large numbers of Muslim traders and the settlement of many Islam-influenced groups, the religion never took general hold in Madagascar, even though there was a settlement of Muslims from Sumatra as late as the fifteenth century. Islam generally left some Arabic words and a few traditions including circumcision, but not much more. Different settler groups retained much of their earlier identity, though all acquired an approximate common language.
In spite of the many gaps in the history of the island between the years before 1500 CE, the Nusantarians were directly, if fitfully, the main agents in the creation of the Malagasy language and nation.
Far Outliers 