Can FIX fix patchy exchange outage communications?

​Picture the scene: paper memos litter the office; a beleaguered manager darts into a meeting room clutching fistfuls of their own hair; a restless employee is glued to a monitor in the corner, feverishly pressing refresh, desperate for good news—any news.

No, this is not a sneak preview of Zack Snyder’s next zombie film…although in their more cynical moments, sell-siders might consider pitching the pandemonium that ensues from a market outage to Hollywood. But the agonizing suspense that makes this comparison possible—the constant, “What happens next?!”—could be mitigated by more standardized communications from venues when markets go down.

Jim Kaye, executive director of the FIX Trading Community, which maintains and develops the Financial Information eXchange messaging protocol, has experienced the upheaval of market outages first hand. Before taking up his role leading the industry standards body, Kaye worked with electronic trading products at Goldman Sachs (2001-2010) and Bank of America (2010-2021), where dealing with venue outage communications fell under his purview. As such, he has been on the receiving end of both good and bad outage communication strategies.

“The first question that goes through your mind during a market outage is, ‘What do I need to do?’” Kaye says.

A FIX working group is trying to help answer this question by designing a common framework for reporting faults. The group is considering the various workflows required by venues and other participants who might need to update their clients on the status of a market, and it plans to design a set of messages that function across all those flows.

“We’re hoping that this FIX work will help speed that process up—not just to get the information out, but also to help decode the message in a business context so people can work out what they need to do,” Kaye says.

The perfect outage

The need for such an initiative is clear. Technology is improving, but market failures are no less frequent: The last year alone has seen high-profile disruptions at Six Swiss Exchange, Nasdaq’s Nordic venues, and Euronext. Each outage causes heavy financial losses and logistical headaches for market participants, but they also erode confidence in the market itself.

There is no single solution to this problem. Monitoring servers and loads, maintenance of hardware, and slick failover procedures all have a part to play. Once an outage has happened, though, good communication is key to keeping markets stable.

As it stands, venues have limited toolkits of varying quality at their disposal to communicate with sell-side trading members in the aftermath of an outage. Most exchanges have an alert on the landing page of their website so that users can quickly find confirmation of the problem in a location that should be unaffected by any bugs or connectivity errors plaguing the exchange.

Next, venues should regularly keep members abreast of the situation. Sakeena Lalljee, head of sales at Aquis Exchange, also leads the relationship function for the exchange’s bank and broker membership base. She liaises regularly with clients in this role, so has a keen understanding of their priorities in the event of an outage. “Best practice is to provide updates at least every 30 to 45 minutes, even if the update is just to say, ‘There is no update,’ or, ‘We’re still working on it.’ The clients value that checkpoint rather than having silence and wondering what’s going on,” Lalljee says.

If it’s too complicated, no one will use it. We need to make sure it’s thorough enough to be useful, but not so thorough that it becomes a burden for everybody

Jim Kaye, FIX Trading Community

The venue might also have additional updates to send, such as how much of the platform is affected (it may be only one order book or one segment of securities), and what steps it is taking (whether trading has been halted, for example, or whether certain functionalities are still running).

This information can be conveyed via machine-readable feeds, which participants can incorporate into their automated trading platforms. More often, though, updates are sent out by email or even phone.

“Most exchanges have a distribution list and some sort of automation or templates where key information just needs to be plugged in. It absolutely shouldn’t be the case that someone is sitting there having to formulate an entire email on the fly,” Lalljee says.

Finally, it is important that the venue notifies trading members that it has resolved the problem well in advance of trading being resumed. This helps with an orderly and confident return to the market. If an exchange has decided to clear the whole order book so that members can start from a clean slate, it is essential that everyone is aware of this arrangement.

“All of this detail matters,” Lalljee says. “Because from a regulatory and compliance standpoint, it’s all about maintaining an orderly market.”

The case for standardization

In practice, however, market outage communications rarely come off without a hitch. When a venue is able to identify a fault in good time, it is unlikely to pinpoint the cause, get systems up and running again, oversee a controlled return to trading, and relay all of this to trading members with plenty of notice.

And the fact that each venue has its own unique playbook and communications channels does not help, either.

This is where standardized outage communications come in. FIX’s working group is still discussing the conceptual design of its project. It is deliberating over what precise pieces of information could benefit from codified standards—a scale to classify the severity of an outage, for example—and what details are best left to the discretion of the venue.

Once it has worked out the details, FIX will begin designing the messages. It plans to have the initiative finalized and signed off by the end of 2023.

In order to win over market players, the system must bring enough standardization to be useful, but not be so prescriptive that it precludes any customization.

“You might want participants forwarding the information to their clients, and perhaps adding some additional information along the way,” Kaye says, adding that sell-side participants may need to prioritize different information in their forward communications. “They’re interested in expressing what it means for their client, rather than just what it means for the venue.”

On the technical side, FIX is doing modeling work to come up with different possible delivery methods. One potential format is the FIX session, a stream of machine-readable messages, which will be easy for most of the industry to ingest given how widespread FIX connections are already. But the communications could also be delivered via web service or even email in a human-readable format.

Noting concerns that FIX could use the initiative to drive further adoption of its own messaging protocol, Kaye emphasizes that the organization is only trying to mastermind a standard language, not dictating how that language would be delivered.

Above all, Kaye believes the system must be simple. This can be achieved by making use of standards (like the ISO standard for identifying venues), reusing existing data designs, and accommodating concepts that market participants are already familiar with, he says.

“If it’s too complicated, no one will use it. We need to make sure it’s thorough enough to be useful, but not so thorough that it becomes a burden for everybody,” Kaye says.

Carrot and stick

Market outage communications may appear to be an inoffensive area in an industry fraught with controversy, but it is still difficult to focus the attention of market participants on the topic.

Iosif Itkin, co-CEO and co-founder of software-testing company Exactpro, has spent the last 14 years trying to reduce the number of software failures at exchanges. He notes that commitment to mitigating outages can be described as a wave function. “After an outage happens, everyone in an organization is very concerned, very dedicated. They take steps to improve resiliency of the system, they introduce additional measures, and the system becomes more resilient for some time. And then people become more relaxed again until the next one,” Itkin says.

The problem of outage communications is so difficult to address partly because it depends on the venue understanding what to communicate, Itkin says. Without a clear idea of what is going on, venues will have a much harder time effectively setting out the next steps.

“The problem turns into disaster when you do not know that you have a problem. If the monitoring is not sufficient, then even a small problem can rapidly escalate because you took the wrong actions or wrong steps,” Itkin says. The unpredictable element of outages makes them very hard to plan for. And while the best response is to increase monitoring of the system using probes, it is easy to see why some venues might find it difficult to tackle such an aleatoric challenge.

But if outages are too infrequent to keep venues motivated, they are beginning to attract the attention of regulators.

The European Securities and Markets Authority (Esma) published an extensive report on market outages in May 2023, outlining the minimum requirements that venues are expected to meet. While it stopped short of prescribing technical standards for market outage communications, it does require that venues publish an outage plan on their websites. In the event of a market outage, Esma suggests that national competent authorities require venues to publish an estimated time of the resumption of trading in the first update and on their websites.

In 2021, the Australian Securities and Investments Commission (Asic) also published a report on the Australian Securities Exchange’s (ASX’s) response to the market outage that resulted from an update to its equities trading platform, ASX Trade. The refresh involved upgrading to the latest version of Nasdaq OMX’s Genium Inet platform, now part of the Nasdaq Financial Framework.

In the report, Asic highlighted its expectation that ASX increase “the content of communications with participants and other market operators” in the event of future outages.

Kaye suggests that having market outage communications in the regulatory spotlight could be a boon for FIX’s standardization initiative. “We’re delighted that the regulators have recognized that this is a serious issue, and that there are solutions to it. Having the regulators’ interest definitely helps bring the industry along,” he says.

But regulatory scrutiny is also heating up the debate. Esma’s report noted that some respondents to its consultation (mainly trading venues) argued that “every trading system is different, and therefore flexibility is necessary to ensure trading venues can resolve incidents and tailor communications in the most effective manner.”

For regulators, the big question here is one of systemic resiliency. While minimizing downtime is desirable, the goal is to ensure that an outage on one market of listing does not cause trading to grind to a halt on all venues. As Aquis’ Lalljee explains, “That shouldn’t happen, because there are more than enough alternative venues in the lit space where trading could continue unaffected, particularly if a primary exchange has a problem intraday.”

In 2021, Aquis and Cboe jointly proposed an industry protocol to mitigate systemic risk in the event of a problem on a primary market affecting either the opening or closing auction. In the report, the exchanges suggest that a group of exchanges could take turns on a rotating basis to function as a backup venue, ready to step in and run the opening and closing auctions when a primary venue goes down. This would help traders to determine a firm price at the beginning and end of each day even in the absence of the venue of listing.

“The more consensus there is on something like this, the better it is for everyone in the market. This is not an area for firms to be trying to get an edge or compete with each other. It’s about trying to have a single best practice that everybody buys into, and which is also realistic,” Lalljee says.

It is unlikely that any single initiative will succeed in making market outage strategies run like clockwork. By their very nature, market failures will always come with a healthy dose of uncertainty. But with threat actors getting more sophisticated and an increasingly erratic climate threatening servers, market participants who hope for the best might be reassured to know that someone out there is planning for the worst.