Back

The promises of the cloud

Positioning to leverage the potential of machine learning and big data

Read the article and learn about:

  • SimCorp’s cloud vision
  • The advantages of software-as-a- service
  • The value of historic data and the potential of machine learning
  • Turning data patterns into insights
The promises of the cloud
Thomas Hejlsberg
CTO, Senior Vice President, SimCorp
Linkedin Connect with Thomas on LinkedIn

In December 2018, Thomas Hejlsberg joined SimCorp as our new CTO. Thomas brings experience from 15 years as CTO / Chief Architect at Microsoft, where he has been part of the transformation of Dynamics 365 Business Central (formerly known as Dynamics NAV) from an on-prem 2-tier solution to a full software-as-a-service (SaaS) cloud offering. Sharing SimCorp’s cloud vision, our CTO explains in this article how we are positioning to service the business models of the future investment management industry.

When a company is moving to the cloud today, the typical driver is to alleviate the cost of infrastructure and instead buy the software as a service (known as SaaS). The argument is that we should buy and use software just like we buy and consume electricity or a streaming service. Nobody would consider having their own tv-studio just to watch tv.

It not only makes sense, moving to a subscription-based SaaS solution also offers many advantages, as opposed to running an on-premise software solution. Among the many benefits are: Automated upgrades, continuous delivery, dynamic scaling, etc. Without a doubt, all these services represent a huge step forward as they allow firms to concentrate on their core business.

SimCorp is already offering to deliver its core system, SimCorp Dimension, as a service, and an increasing number of our clients are utilizing this option. If you have not made the move to the cloud, I can only recommend starting your transition today, rather than tomorrow. If you are not yet convinced, keep on reading.

This is only the beginning

The true promise of the cloud is unleashed when we start combining services. You might say that we can make 1+1 become 3. But it doesn’t stop there. Machine learning (ML) and artificial intelligence (AI) will be able to provide us with something additional: “Insights” – and this is where there is no limit to how much 1+1 can add up to.

Machine learning

To understand insights, however, we first need to get a general understanding of ML. Let’s start with what is greatest about it: There is no magic involved and it is easily understood. Just like we can watch a movie without knowing exactly what it takes to make one, you can say the same thing about ML.

The core capability of ML is the ability to predict something based on existing (historic) data.

We do it ourselves all the time. If a train has been on time 99 times in a row, we predict that it will be on time the 100th time as well. If our teenage son has overslept five out of the last ten days, we would not expect him to be on time the next ten days, suddenly making a perfect attendance. Rather, we would predict that he will not make it.

Try to translate this reasoning into a common business scenario. Can we trust a given customer with a line of credit? Well, by help of our historic data, we can find out whether other customers within the same line of business, having been in business approximately the same amount of time, having a comparable revenue, etc., made it. Based on this data, we can predict the risk of our new customer defaulting payments.

The ML functionality simply finds patterns in the data matching the facts of historic data and applies the pattern to the new data to form a prediction. Thomas Hejlsberg, CTO, Senior Vice President, SimCorp

 

Combining data to form patterns

You might ask, is it really that easy? And the true answer is that finding the patterns that leads to trustworthy predictions (we call that a model) is quite difficult and requires a lot of complex math and computer power. But this is luckily not our concern - or yours. What matters is to be able to use ML and thereby the patterns that have proved to work, which does not require special math skills nor significant computer power.

Not that long ago, I got a notification on my phone: “You need to leave now, to make it home to your dinner appointment at 6 pm.” Nice… But, hey wait… I actually never told my phone where I live, where I am, my preferred way of transportation, nor anything else.

What happened here is that the phone started combining data and historic facts. The place I consistently sleep at night is in all likelihood “home”. My position right now can be acquired from the GPS in the phone. The fact that I have guests coming is in my calendar, and also that the location is “home”. The traffic situation can be picked up from a service… and since I drove my car this morning… you get the picture.

And this is where the “insight” is born. I did not have this insight that due to an accident on the highway, I needed to leave early.

The system actually told me something I did not know,
and I didn’t even ask for it.Thomas Hejlsberg, CTO, Senior Vice President, SimCorp

This is the core of insights. Information or recommendations that help you and that you either wouldn’t have thought of, or data that would have been very difficult to filter out from the vast amount of information hitting us every day.

You can keep combining data from various services, e.g. combine sales, customers and sales people with LinkedIn, and you can predict which schools/educations yield the best sales people.

Combining business data and social media data enables mind-blowing scenarios. And this – in my opinion – is the second, and often overlooked, reason to move to the cloud.Thomas Hejlsberg, CTO, Senior Vice President, SimCorp

Fulfilling the promises of the cloud with big data

All this becomes possible when we combine multiple datasets from multiple services. This is known as big data, another term typically found near ML.

Combining different data sources (big data) and applying ML to this data to form new unknown insights, has the potential of being a complete game changer. “Wow, this is going to change the world,” you might think, and I would have to agree with you.

I also believe the change will be for the better. ML will in a not so distant future be able to predict how to best fight diseases, predict adverse weather or do natural live speech translation, just to mention a few areas. Very few people or businesses will remain unaffected by this change.

In fact, I predict (without use of ML), we will see many changes that we all can benefit from. New business models, new offerings will emerge.

“Will SimCorp invest in ML and what is the perspective?” people often ask me. Yes, absolutely. ML is without doubt a vital part of SimCorp’s strategy. Some time ago, we set up SimCorp Technology Labs, with the sole purpose of investigating the concrete ML scenarios within SimCorp’s reach. We have some interesting ones in the pipeline, which will provide our clients with valuable insights. Furthermore, allowing our partners and clients to combine services is where the innovation can scale.

We are currently investing significant resources in a scalable platform (SimCorp Evolution) that allows highly scalable workloads to be added to the already highly capable SimCorp Dimension platform. ML is a natural part of such an expansion.

SimCorp’s cloud strategy is ambitious

Another part of our strategy is to offer capabilities that will allow workloads to offer data as a service, subscribe to other services, as well as ML implementations – currently all these areas are ongoing or planned activities. Through our Technology Labs, we have a close co-operation with universities, where we offer candidates to finish their education in cooperation with SimCorp, a win-win for multiple reasons.

Since all ML predictions are based on data, data itself has become a commodity. Major players like Amazon, Google, IBM, Facebook or Microsoft are all competing to create the largest pool of “big data”.

This situation poses questions like: Who owns the data? Who owns the models (patterns derived from the data)? As members of the software industry, we must be able to respond to these questions. GDPR and other initiatives worldwide are designed to frame this process. Obviously, we prioritize security and adherence to legislation above everything at SimCorp. Also, new business models will evolve. SimCorp is closely monitoring this development.

Let me finish by stating: I see all this as a journey, and this is going to be a very interesting one of its kind. This was one of the main drivers for me accepting the challenge to take on the position as CTO at SimCorp – another driver being the people in the organization (and their passion). I simply can’t help being excited by the potential of both. Stay tuned, I look forward to sharing more updates from our cloud journey.

Share this story