Chief Technology Officers and IT architects of a certain age will recall several 'pivotal choices'. These choices aroused such great passions that they felt more like religious wars than businesses trying to work out what was good for them. Mainframe vs network, Mac vs PC, Windows vs Unix, Oracle vs SQL Server, Java vs C#. Another such battle looms today with cloud providers Amazon AWS vs Microsoft Azure vs Google Cloud.
Occasionally it was important to get those choices right. In the early nineties, some large companies over-valued their relationship with IBM, the goliath of the industry, and put their money into IBM hardware, a token-ring network, and OS/2. That proved to be the modern-day equivalent of joining the dinosaurs the day before the meteorite hit.
Other pivotal changes were not the straight one-or-the-other choices they appeared to be. Many companies thrived for decades despite running a multitude of operating systems and database platforms. There would be occasional grumbles at the seeming wastefulness of duplicated resources and know-how. But ultimately the benefit of consolidating down to a single platform did not clearly justify the cost of switching. This pluralistic approach also had the benefit of allowing a kind of competitive natural selection to take place, and it avoided the calamity of central planning mistakes.
So what type of pivotal change is cloud going to be? Is it a fork in the road, or is a multi-cloud strategy a viable route? What considerations are pertinent for investment management companies?
BlackRock recently announced its selection of Microsoft as partner to shift its Aladdin platform to the public cloud. This follows a similar announcement last September by Aladdin's rival, Charles River Development (CRD), concerning a partnership with Microsoft. Previously, both firms had spoken only of private datacentres and cloud-based hosting for their systems, presumably as comfort to those clients having a deep mistrust of public clouds and mistakenly associating the word 'public' with 'insecure'.
The announcements provide yet more evidence that the major cloud providers have been extraordinarily successful in commoditizing the datacentre business. They run datacentres and host applications better and cheaper than anyone. There really is no virtue left in saying: 'we run our own datacentres'.
CRD's choice of a Microsoft platform to host a Microsoft-based application seems natural enough but, to some, BlackRock's choice is surprising. The Aladdin development teams using non-Microsoft technologies (Java and Perl) greatly outnumber those who develop using Microsoft-native technologies. With Amazon and Google being more rooted in the Linux/Java world, wouldn't one of those vendors have made more sense?
In fact, the assumption Windows therefore Azure, Linux therefore AWS doesn't hold at all. Amazon boasts that AWS runs more than twice the number of Windows Server instances as Azure. Correspondingly, Microsoft is rather proud that that Linux now dominates Azure's server VM market. So we shouldn't confuse a vendor's cloud capabilities with the development technologies associated with the same vendor. The top two cloud vendors basically do a good job of hosting any workload.
The Aladdin and CRD announcements are concerned with moving their respective application infrastructures into public cloud datacentres. In both cases, the architecture was established twenty or more years ago and has evolved only incrementally. Their 'moves' signal a change in where the services are hosted, but not a fundamental re-architecting of those services.
What could be better than merely hosting in the cloud? An application architecture that is designed natively for the cloud and that takes full advantage of the opportunities cloud computing presents. That is exactly the approach that Ryedale has taken with its cloud-native investment platform.
Ryedale chose Microsoft Azure as its home cloud from the very beginning it was founded in 2012.
Our goal was rather different to that of Aladdin and CRD. We aimed to develop the next generation of investment system and to make it cloud-native from the outset. As software developers, we wanted clear, well-defined APIs that fitted together well to build a big system. Azure has delivered that for us even as cloud-computing has evolved and transformed enormously in the intervening years.
Sean Kelly, co-founder and Chief Technology Officer, Ryedale
So, bearing in mind all the above, you might expect us to say that Azure is clearly the leader for investment management businesses. However, we feel strongly that multi-cloud is the way to go.
This is one of those pivotal choices where the outcome for a firm is better if there is no internal monopoly. The very fact that cloud provides diverse, flexible, pay-for-what-you-use computing resources over the internet means not just that you can use the services of more than one vendor - it means you should.
You can be pretty sure that Azure does not have an internal monopoly within BlackRock or CRD's parent State Street. It's probable that teams of developers and/or researchers there routinely use AWS, even if it is not the home choice of the firm. For clients whose main concern is consuming software application services, the whole point of SaaS is that you don't have to care about the hosting technology. The choice of cloud makes no difference to consuming an application, provided the application works as it should.
We understand cloud better than most investment management platforms. That's why we're committed to multi-cloud.