The financial markets often experience bouts of volatility, but market activity since the start of the pandemic in 2020 has been unprecedented. The economic turmoil sparked by the global pandemic created new processing challenges throughout the trading lifecycle, pressuring market participants’ capabilities in the 帖子-trade space. 存的白皮书，”Managing Through a Pandemic: The Impact of COVID-19 on Capital Markets Operations,"发现在, and in the immediate aftermath of the COVID-19 pandemic, 该行业保持弹性. However, opportunities remain for further optimizing 帖子-trade processes across the capital markets.
哈罗德Watler, 存董事总经理, 企业平台工程, 和Mark Cucarese, 存执行董事, 企业生产服务, met with 存管连接 to discuss how IT stays on top of spikes in volatility in order to ensure our systems remain stable and operational as the financial markets meet new challenges.
Related: Cyber Risk and 操作弹性 Podcast
DC: What factors have led to the spikes in trading activity over the past two years?
MC: A combination of factors fueled remarkable spikes in trading activity, including an influx of new market investors and the unexpected resurgence of Covid-19, 是什么阻碍了经济复苏. With millions of new investors and the rapid rebound after the March 2020 selloff, 市场成交量处于历史水平.
DC: How does 存 help clients handle market volatility?
MC: Our 企业生产服务 team provides 24/7 support for the entire 存 global infrastructure, 包括所有主机, 分布式平台和基于云的系统, with a central focus on production reliability for the financial markets. We’ve created a virtual command center that monitors our production systems, 检测任何中断, 并实时响应警报. We have hundreds of technical support staff around the globe constantly engaging with colleagues and clients to proactively investigate any market anomalies.
Our team builds tools to support systems monitoring and event response. 例如, as part of ongoing IT modernization we created a tool to help us understand Universal Trade Capture performance, with visibility into peak activity or where a trade is in the processing flow. During a period of heightened market volatility we worked closely with one of our clients, and were able to understand their transaction patterns and how to improve their performance by using this tool.
DC: 平台工程在这个过程中的角色是什么?
HW: 在我们的支持结构中, Enterprise Production has the first level response, even the second-level triaging of any issues and then they escalate it to Engineering. We join forces for other production environment issues or problems.
我们设计, plan and maintain 存’s mainframe and private cloud footprint, 以及我们的网络能力. Capacity is very important, and we’re always keeping an eye on trends. At the end of 2019 we started to see trading volumes increase. 然后Covid-19来了. 很明显，这不是一个异常现象, so we managed with all the horsepower we could muster to keep things running smoothly.
DC: How did Platform Engineering adjust its strategy to account for extreme events?
HW: As 2021 began, we expected the heightened market volume and volatility to continue. We examined the resources we would need to sustain that level and beyond and the cost would have been astronomical. So, we accelerated our 2022 mainframe refresh, which included increasing our processing power by 35%.
We now have infrastructure whose performance gives us a lot more capacity overhead. Even on very busy days we’ve been able to scale up as needed with the capacity we have immediately available. I’ve given up thinking anything will ‘calm down’ anytime soon.
DC: How does auto-scaling ensure we can respond to market volatility?
HW: Auto-scaling is the ability to dynamically add or remove resources as needed. It’s a prepackaged capability that we utilize to handle spikes for some of our products.
在我看来, 如果你看一下需求容量, when you have the additional resources on-premises and readily available, 我们不需要获得这种能力.
Self-healing is where we develop this capability ourselves, 我们有工作和剧本说, 如果满足这些条件, 添加另一个引擎.’
DC: What is being done to enhance production monitoring for the future?
MC:我们有一个强有力的监控程序, and we’re aggressively accelerating automation efforts, 建立增强的健康检查, 仪表板和工具. We’re eliminating manual touch points and building self-healing systems to automatically respond to alerts - before an incident can occur.
今年, we launched the first implementation of our artificial intelligence operations tool for incident prevention. It uses predictive analysis based on historical trends to respond to alerting and predict whether something is going to become an issue. Then the team can take preventive measures to avoid a production outage. Think of it like pilots in a cockpit, always managing the gauges.