2023年4月25日

大雪橇了17美元系列来简化云间数据共享大雪橇,创业就是构建一个云间数据共享平台,今天宣布它已经筹集了1700万美元资金轮系列共同Greycroft和Madrona企业集团。这一轮之前700万美元的种子轮由.406企业。在这最新一轮,大雪橇的估值是8700万美元。大雪橇在2021年末成立之前,该公司的首席执行官和共同创始人杰克•格雷厄姆是一个主要产品经理在微软Azure在Azure数据交换。格雷厄姆,他也曾在Neo4j和英特尔,认为数据共享是一个基本分析景观的变化。“这是下一个进化的企业如何获得他们需要的数据驱动机器学习分析,数据科学,你,”他说。“我已经建立在数据基础设施的空间大约十年,api和SFTP的数据移动只是其中之一,我接受。直到我真正开始挖掘我是:它不应该。这个没有很大的意义。这些都是为以前的模式构建的数据。” Image Credits: Bobsled Graham also believes that it’s important for the clouds and data platform to create the building blocks of what a native sharing protocol should look like. “If you believe that data sharing is the way the world is going to go and that there’s going to be much more active collaboration on data between source and destination, you kind of need there to be something that brings all of these different platforms together, because there’s not one right cloud, there’s not one right data warehouse, there’s not one right layer of the stack for it to make data sharing happen,” he said. Bobsled, unsurprisingly, believes that it can be this neural, third-party data sharing platform that can allow businesses to connect their various data sources in a more native way. The service uses each platform’s sharing protocol and then connects the various sources and helps businesses prepare the datasets for querying them. Image Credits: Bobsled Some of the common use cases here are SaaS applications that may want to share application data from AWS S3 into a customer’s Azure storage, for example, or a data team that wants to move sales data from Azure to S3 for analysis. Obviously, there are other companies that focus on this kind of data transfer and data transformation space. But Graham believes that with these ELT platforms like Fivetran, the consumer still has to do too much work. “A lot of what actually is being done by ELT platforms should be done by the source of the data,” he said. “So we’re shifting left a lot of that. If you pull apart the letters, extract: you’re doing that because it’s not being shared with you. Load? Same thing. If you’re getting a share within Databricks, you’re not loading into anything.” As for transforming data, he noted that most of the time, the data is in a transactional data model that needs to be formatted for analysis and the source should transform that data — though he believes that transforming the data from multiple data sources to match a given scheme is always going to be up to the consumer. “I wouldn’t call us directly an ELT product yet because we’re different enough, but I think that’s the closest thing to the category that we’re building. It’s just a different user,” Graham explained. The company plans to use the new funding to build out its platform, mostly with a focus on support for more data sources (including old-school SFTP sources). “The promise of AI only heightens the urgency and increases the priority for companies to realize the value of their data,” said Madrona Managing Director S. Somasegar. “Even within enterprises and amongst partners, that data is now spread across platforms not designed to work together. Bobsled‘s unique cross-cloud data sharing focus will unlock the value of data that has been complex, slow and expensive to access until now.”