腾讯云窦贤明：单一数据库不能解决所有问题

{"type":"doc","content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在数据量持续爆增、数据日益多样化的今天，传统数据库的迭代速度已经追不上数据的增速，且企业对数据库计算和存储能力的要求越来越高。面对当前的挑战和机遇，国产数据库厂商的研发创新速度不断加快，可以说云计算时代的到来，扭转了国外商业数据库一家独大的局面。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"目前，国产数据库领域正处于百花齐放的状态，已经有越来越多的行业巨头参与到了数据库的建设中，腾讯云便是其中之一。为了更深入地了解"},{"type":"link","attrs":{"href":"https:\/\/cloud.tencent.com\/?fromSource=gwzcw.2212127.2212127.2212127&utm_medium=cpd&utm_id=gwzcw.2212127.2212127.2212127","title":"xxx","type":null},"content":[{"type":"text","text":"腾讯云数据库"}]},{"type":"text","text":"的发展历程，从而进一步透视国产数据库的发展方向，InfoQ 和腾讯云数据库专家工程师窦贤明聊了聊云数据库的发展、前景与挑战。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"单一数据库不能解决所有问题"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"进入基础软件领域已有十余年光景，窦贤明亲历了云数据库从零到一的建设过程，作为整个浪潮的参与者和见证者，他对技术、产品以及市场都有着更加深刻的认识。据他介绍，当开发人员在部署一个传统数据库时，需要涉及购买硬件、部署机房、建立网络、部署实例、规划资源等等一系列操作；在维护传统数据库时，还需要进行扩容、监控、告警、日志、参数设置等等操作，而云数据库的出现便能够更加轻松、简单地实现上述工作。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"开发人员可以直接在云数据库控制台完成数据库的申请和创建，几分钟内便能准备就绪、投入使用，通过云数据库提供的控制台，还可以对所有实例进行统一管理。云数据库还支持物理备份、逻辑备份、备份恢复及秒级回档等功能，以此来保障数据的安全性。此外，传统数据库的价格高昂，动辄就需要投入数十万元的成本采购设备，而云数据库则能够按需付费，用多少付多少。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"尽管相较于传统数据库，云数据库已经能够帮助企业解决大部分问题，但窦贤明告诉 InfoQ：“单一数据库不可能解决所有问题，云数据库在存储成本、HA 切换、网络瓶颈方面依然存在优化的空间。”"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"如下图所示，Master 和 RO 虽然对应的是同一份数据，但在存储上实际有六份数据；而每多加一个 RO 节点就会多出三份数据，也使得整个集群的存储副本数近一步放大；高吞吐的数量会使网络问题成为瓶颈，在共享存储侧也有大量网络浪费。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/f9\/f9d19d0560dcdcae046a7317aaf85cea.webp","alt":"图片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"云原生数据库应运而生"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"目前，窦贤明与他团队研发的"},{"type":"link","attrs":{"href":"https:\/\/cloud.tencent.com\/product\/cynosdb","title":"xxx","type":null},"content":[{"type":"text","text":"云原生 MySQL"}]},{"type":"text","text":"（TDSQL-C MySQL）、云原生 PostgreSQL（TDSQL-C PostgreSQL）便能够很好地解决存储成本、弹性扩容等问题。作为新一代企业级云原生分布式数据库，它的初衷是为了让运维人员更省心，让数据库的运维变得简单，具体来说，TDSQL-C 有以下产品特点："}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"bulletedlist","content":[{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"全面兼容：100% 兼容开源数据库引擎 MySQL 5.7 和 8.0 以及 PostgreSQL 10。几乎无需改动代码，即可完成现有数据库的查询、应用和工具平滑迁移。"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"高性能：具有商用数据库的强劲性能， 最高性能是 MySQL 数据库八倍、PostgreSQL 数据库的四倍。"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"海量存储：最高 128TB 的海量存储，无服务器 Serverless 架构，自动扩缩容，轻松应对业务数据量动态变化和持续增长。"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"快速恢复：计算节点实现无状态，支持本地和跨设备的秒级故障切换和恢复；支持基于快照的秒级备份和回档。"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"数据高可靠：集群支持安全组和 VPC 网络隔离。自动维护数据和备份的多个副本，保障数据安全可靠，可靠性达 99.9999999%。"}]}]},{"type":"listitem","attrs":{"listStyle":null},"content":[{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"弹性扩展：计算节点可根据业务需要快速升降配，秒级完成扩容，结合弹性存储，实现计算资源的成本最优。"}]}]}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"TDSQL-C MySQL \/ PostgreSQL 基于共享存储实现了存算分离架构，Master 和 RO 是基于一份数据放在共享存储中，RO 只从共享存储中读取所需的 page，不需要写入存储，并且 RO 可以从主库接收 WAL 在缓存中重放，以此保持缓存中 Page 持续更新。这样一来，云原生数据库便解决了业务容量和计算节点的扩容的问题。TDSQL-C 还能够自动判断计算层面的资源，实现计算关停和热启动，且启动时间大概在 3s 以内即可完成。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"image","attrs":{"src":"https:\/\/static001.geekbang.org\/infoq\/e5\/e53c16e070e03deddcc8c60b1c916753.webp","alt":"图片","title":null,"style":[{"key":"width","value":"75%"},{"key":"bordertype","value":"none"}],"href":null,"fromPaste":true,"pastePass":true}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在 11 月 4 日的腾讯数字生态大会 Techo Day 上，腾讯云副总裁李纲还宣布了云原生数据库 TDSQL-C 的全新升级：吞吐率提升 50%、将 IO 延迟降低 80%，整体性能提升 85%；带来全新形态 Serverless，通过全局工作流预测以及动态扩缩资源，进一步降低成本，做到真正的按需计费。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"攻克最严苛的领域"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在窦贤明看来，现阶段云数据库能够解决企业怎样的问题想必已经没有分歧，但对于一些传统企业来说，毕竟要将自己所有的业务数据迁移到公有云上，安全性成了他们最大的顾虑。关于这一点，他表示：“扭转行业内的固有认知是当下的一大挑战，云是一门信任的生意，需要长期积累的过程才能扭转这样的局面。”"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"那么，对于腾讯云数据库来说，怎么做才能加速对行业的渗透？“只要攻克了最为严苛的领域，就能证明我们可以满足绝大多数企业的需求。”窦贤明给出了这样的答案。而由于金融领域本身的业务特点，使其在数据一致性、高可用，性能成本以及水平伸缩等方面都有非常严苛的要求，也因此，金融领域成为腾讯云数据库势必要攻克的一块高地。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"腾讯云副总裁李纲也曾公开表示：“国产数据库的发展一般会经过互联网企业、民生政务、传统行业应用、金融核心业务这几个阶段的打磨，其中金融行业对数据库要求最为苛刻，不仅数据容错度低，而且还要符合信息安全等级规范。”"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"作为国产分布式数据库的重磅产品，"},{"type":"link","attrs":{"href":"https:\/\/cloud.tencent.com\/product\/tddbms","title":"xxx","type":null},"content":[{"type":"text","text":"TDSQL "}]},{"type":"text","text":"在背后支撑了全国第七次人口普查、防疫健康码、张家港农商行核心系统的落地应用等等，且服务了国内前 10 大银行中的 6 家；在政务、电信运营商等领域，也已经服务了超过 3000 家金融政企客户。此外，在保险行业，TDSQL 正在帮助太平洋保险集团实现全面数据库国产化。在这个项目中，TDSQL 的 Oracle 兼容性得到了充分验证，兼容度高达 98% 以上，这能帮助业务在极短时间内，极小业务改动量的情况下，快速完成测试验证和上线。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"在采访过程中，窦贤明还为我们介绍了一个“微服务 + 国产分布式数据库”的架构案例：昆山农商行将新核心系统划分成公共服务微服务集群、账务微服务集群和历史微服务集群，并把这三个微服务集群运行在一套 TDSQL 集群中。由于 TDSQL 保障了微服务间一致性读的问题，使得企业在应用微服务组织结构的同时，也能解决存储分布式扩容的问题。"}]},{"type":"heading","attrs":{"align":null,"level":2},"content":[{"type":"text","text":"持续迭代数据库的五个基本能力"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"纵观腾讯云数据库产品家族，包含了关系型数据库、非关系型数据库以及数据库迁移、智能运维、可视化平台等相关应用。除了兼容主流的 MySQL、MariaDB、Redis、Mongo 等开源技术，腾讯云数据库也有内部自研的产品分布式数据库 TDSQL 和云原生数据库 TDSQL-C，基于这样多点开花的局面，也让我们不禁产生了一个疑问：面向未来，腾讯云数据库重点突围的方向会是什么？"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"“稳定、安全、易用，高效、成本低”窦贤明这样告诉 InfoQ，由于基础软件与应用软件在迭代速度上有着很大的不同，所以它不可能一天一个新概念。在他看来，做数据库要学会坐冷板凳，需要朝以上五个基本能力持续演进、迭代并逐步做到极致。窦贤明介绍说：“当前我们已经做到了 99.95%，未来需要朝 99.99%，甚至更高的目标去努力。”因为当云数据库没有了稳定性、可靠性、安全性等最基本的东西，一切都将成为空谈，也势必会被市场所抛弃。"}]},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null}},{"type":"paragraph","attrs":{"indent":0,"number":0,"align":null,"origin":null},"content":[{"type":"text","text":"“在接下来的 5~10 年里，国内数据库行业将会出现一个很大的变化，相信过不了多久，大家会认可国产数据库是更好的数据库。”而这句话背后，不仅仅代表了腾讯云数据库的信心，更体现出了国产数据库从业者的底气与实力。"}]}]}