跳到主要内容

How 存 Achieves Data Resiliency with Snowflake’s Snowgrid Technology and AWS

Arjun宫著, 存总监,Snowflake和数据平台团队负责人 and Sudha Gullapalli, 存数据云工程副总监 | 5 minute read | October 10, 2023

业务连续性仍然是全球公司的首要任务, 考虑到自然灾害造成的破坏, 区域网络和停电, 网络攻击和破坏, 用户错误(仅举几例)不是问题 if 但是一个 .

Arjun宫Arjun宫, 存总监,Snowflake和数据平台团队负责人

The case for business continuity is particularly compelling for a company such as The Depository Trust & 结算公司(存), which is designated as a systemically important financial market utility (SIFMU), a U.S. Congress-enacted status recognizing that disruption or failure of such an organization would destabilize financial markets. This is why 存 is committed to delivering the world’s most efficient and resilient 帖子-trade financial market infrastructure. Snowflake on AWS supports our business resiliency initiatives and enables us to meet and scale disaster recovery with operational efficiency and confidence.

Sudha GullapalliSudha Gullapalli, 存数据云工程副总监

在我们继续Snowflake和AWS的故事之前, 这里有更多vns6060威尼斯城官网存的信息,以帮助您了解其中的利害关系. 我们结算了美国的大部分证券交易.S., 4美元.每天5万亿美元.S. 政府债券和每月平均$8.35万亿的抵押贷款支持证券. You get the idea: business continuity is imperative for us for settling securities transactions or to run internal reports, so our IT strategy is based on the three foundational pillars of security, 弹性和稳定性.

使用Snowgrid在每个元素中构建弹性

在存, 我们所有的计划都包含了弹性的概念, whether for clearing securities or offering client’s the ability to perform data analytics, 包括如何使我们的应用程序现代化. 每个应用程序都有一个灾难恢复计划, 包括我们所说的手册, detailing the failover and failback schema as well as the objectives for the two main criteria in disaster recovery:

  • 恢复点目标(RPO): 如果数据丢失,您期望实现的数据恢复范围.
  • 恢复时间目标(RTO): The maximum amount of time you will tolerate an application not being available in the event of disaster.

Since implementing Snowflake on AWS in June 2020 for our risk and data analytics, 我们的组织没有发生过事故. 这种弹性成功的原因之一是雪花 Snowgrid 功能. Snowgrid enables customers to replicate data and more across regions and clouds, 释放更大的弹性并最大限度地减少业务中断.

We have conducted at least 15 disaster recovery exercises using Snowgrid technology for business continuity. 我们的Snowflake实例处理超过700个,每天在15个应用程序中查询000次,支持400多名用户, and we have been able to achieve close to zero data loss and near-zero RTO using Snowflake’s account 复制 功能.

Snowflake’s built-in redundancy is a major benefit for 存; there is triple redundancy for all critical services and automatic retries for failed parts of any query. 在区域层面, Snowflake uses availability zones on AWS and also offers cross-region 复制 和故障转移, which has helped us achieve our business continuity goals of close to zero data loss and near-zero recovery time objectives. We can use the Snowflake Time Travel feature to query for and retrieve deleted data for up to 90 days—and a fail-safe feature offers an additional seven days past the retention period for time travel.

Snowgrid’s account 复制 功能 allow each account to have one or more failover groups, 所以我们可以根据业务线来区分应用程序. 这为我们的灾难恢复过程设计提供了很大的灵活性, including the ability to fail over an app with its own connection URL intact, so the app and its connection fail over together (and can fail back together as well). We also gain the ability to rotate apps independently without impacting one another.

收获好处

我们总是力争RTO为零. Snowflake用它的许多关键特性支持这种努力, 包括多云支持, 随需应变的可伸缩性, SOC 1和SOC 2合规性, 复制, 和故障转移. 在过去的9个多月里,我们做了弹性(混乱)测试, 压力测试, and testing P99 lags; we feel that we’ve put Snowflake 复制 through thorough testing and had good success.

在存, Snowgrid复制和故障转移的好处包括一致性, 速度和成本节约.

一致性: 自动化d syncing across primary and secondary accounts and cloud providers eliminates manual migration tasks for operational efficiency. Each application has one runbook for disaster recovery (DR) processes globally, meaning there is only one code base for centralized management and execution of 复制. 我们可以为美国使用相同的代码库和过程.S. 和欧盟,省力.

速度:可启用容灾, tested and equipped with its runbook (detailing DR plans) in less than three days. The simplicity and elegance of design make it fast to work with Snowflake for DR.

节约成本雪花复制是便宜的. Our previous on-premises 复制 solution doubled our cost due to having to double the hardware and licensing costs.

用雪花的计算和存储分离, 高度压缩的微分区被复制, which improves storage efficiency and data freshness at the 复制 site. 与即时启动计算vns6060威尼斯城官网的能力相结合, we are able to recover quickly while paying for the compute only 当 needed. Avoiding the need to dual-load and transfer data (ETL) has helped us realize savings of roughly 30%.

在存, 我们所有的计划都包含了弹性的概念, 包括如何使我们的应用程序现代化.

业务持续成功的四个秘诀

存’s partnership with the Snowflake team gives us a close and constant feedback loop and the opportunity to try out new features while in private preview. 在一起, we’ve made it possible to move big rocks—complex things like System for Cross-domain Identity Management (SCIM) provisioning and user 复制.

当您承担(或继续)您自己的业务连续性计划时, 我们强烈推荐Snowflake作为基础,并提供以下建议:

  • Make sure you understand your company’s assets and identify what represents acceptable loss or downtime (if any) for each application.
  • 不断地测试并寻找边缘情况.
  • 自动化, 自动化, 自动化—it’s the only way to achieve the scale and efficiency needed for mission-critical applications.
  • 保持测量以持续改进.

在存, we pride ourselves on designing our IT strategies for resiliency right from the start.

与雪花, 以及Snowgrid的跨云能力, we know that the security and operations aspects of our architecture are covered so we can focus on optimizing the user experience and adding value to our business.

对Snowgrid感到好奇? 读了 使用Snowgrid在全球范围内操作 解决方案简单.

帖子
存管连接
2023年9月21日 释放独特的潜能...
帖子
存管连接
2023年9月25日 开启数据流动性的未来...
帖子
播客
2023年10月3日 提高结算效率...
回到存连接
dtccdotcom