The Sulzer Group is fully committed to cloud technology. What clinched your decision to adopt it?
Urs Schmied: There used to be about 3,200 servers in service in the Sulzer Group around the globe. When the time came to renew a large portion of this hardware in 2013, we asked ourselves whether this investment was really making sense.
Evidently you weren’t convinced that this was the case...
Schmied: No, we weren’t. Although cloud technology has evolved in great leaps and bounds over the last few years, back then it was already rather obvious that it made little sense to operate our own servers all around the world. The harmonization and consolidation of our Group IT was already well underway at that time. So we asked ourselves the fundamental question: Where do we want to go with our IT?
Sulzer, headquartered since 1834 in Winterthur, Switzerland, specializes in pumping solutions, services for rotating equipment as well as in separation, mixing and application technology. The company creates reliable and sustainable solutions for its markets: oil and gas, power, water, and general industry. Sulzer serves customers around the world through a network of over 180 production and service sites in 40 countries, with 44% of the orders coming from emerging markets. In 2016, the company achieved sales of roughly CHF 2.9 billion with around 14,000 employees.
And what was your conclusion?
Schmied: One of the Sulzer Group’s goals is to build the best pumps in the world. IT is a means to an end. So we want to concentrate on providing functionality for the business. We’re not interested in maintaining hardware for its own sake.
A lot of companies would like to follow the path taken by Sulzer. But they’re plagued by general uncertainty. What would you see as the most important success factors?
Schmied: To start with, you need to define a precise goal that is tailored to the company’s requirements and strategic aims. The goal can naturally be a bit ambitious, but it must be something you can actually achieve. Then to flesh out the vision, you need the right people with intimate knowledge of the company, who know what is possible in terms of IT.
"What ultimately counts is how much it costs overall. And today that figure is significantly lower, when all’s said and done. With the size of our operations, we’re talking about several million francs."
Urs Schmied, Head IT Infrastructure
In your case, where did the know-how come from? Perhaps from external consultants?
Schmied: No, for this type of radical transformation you need to have the know-how within the company. We were lucky in that the new CIO, Ursula Soritsch-Renier, joined Sulzer around that time. After we had developed the new IT strategy under her leadership, we built an internal team of cloud experts.
Then came the next and probably the most important step: we had to convince the management of this vision. For us it was clear that it wouldn’t work without the full support of top management.
Why was this step so important?
Schmied: When you tackle this kind of migration project, you are bound to encounter problems in the implementation phase. If you have the entire management against you at the very first obstacle, things can get very difficult. That’s why discussing it beforehand with senior management is extremely important. For Ursula Soritsch-Renier it was essential to radically transform the role of IT in the Sulzer Group – and with it the perception of IT in senior management. Previously IT had always been seen merely as a cost factor. We wanted IT to become a partner to the business on an equal footing.
And how did you manage to convince the management?
Schmied: As in all organizations, it was a process. First, we focused on building a common understanding of what the cloud was – and what it wasn’t. The cloud was mostly seen as something negative. There were many questions to be answered, particularly about risks in the areas of data availability and security, of course. But we soon managed to convince the management that the cloud was in fact already playing a role in our organization. SAP and applications in HR were already being provided as cloud services. The experiences gained there were thoroughly positive. Naturally that helped us enormously.
So, by far the most important thing was to reach a joint understanding of what «the cloud» was and how we intended to deploy this technology going forward. To do this, we had to make our visions tangible in the form of concrete plans. This also helped us gain the confidence of the management. Our plan was founded on specific examples of how we would proceed in the transition.
How were you able to address the issue of security?
Schmied: Security was a key issue for us right from the word go. Which is why Andrea Klaes, our CISO (Chief Information Security Officer), was involved in our deliberations from the start. Close collaboration with Security, Compliance and Legal departments is especially important when it comes to the cloud. Our plans didn’t just describe the possibilities of the cloud but also the challenges and the risks. That’s a key issue: in my opinion, part of preparing professionally is understanding exactly where the potential sticking points are.
Your vision is Cloud First. What exactly does that mean?
Schmied: We have a clear goal: long-term we no longer want any hardware of our own. The implementation works like this: whenever we identify a requirement in consultation with the business, first we examine whether there’s a SaaS (software as a service) that provides this functionality. If not, we look at PaaS (platform as a service) or IaaS (infrastructure as a service). Only if the use of cloud technology is really impossible – for legal reasons, for example − do we go or stay on-prem, that is, on our own physical hardware on premises.
You had quite an ambitious time frame...
Schmied: In 2013 we elaborated the vision and our plans were endorsed by management; in 2014 we launched the two proofs of concept (PoCs). Then in March 2015 we put it out to tender. Due to the intense work entailed by the strategic framework and planning, as well as the experience gained from the two PoCs, the team could define the requirements to an external partner very precisely and link them to concrete operational tasks. The biggest challenge consisted of migrating an existing data center in Stockholm – with 100 servers and about 1,200 users – to IaaS. This concrete case enabled us to make a precise analysis of the offers, above all with regard to the costs.
In September 2015 we then decided on two partners, CenturyLink for IaaS and PaaS, and Microsoft for SaaS and IaaS. This was followed by the migration itself. Fifteen months later – at the end of 2016 − we reached our goal and switched off the lights at the last data center.
Today you live in a hybrid cloud environment. What does the current distribution look like?
Schmied: To save time during the migration, in many places we adopted the «lift and shift» approach. That is, we migrated most applications and data directly to IaaS. Which is why today we use IaaS a good 80% of the time. We cover about 5% with SaaS, which includes Office 365.
The proportion of SaaS is sure to grow further in time. The remaining 15% are on-prem as before, particularly at our production locations. Although it’s a downwards trend, at certain locations there will probably always be servers running local applications.
What would you say are the advantages that you have now?
Schmied: Our infrastructure is secure and stable, but at the same time it has become much more flexible. This naturally benefits the business directly. We are now much more agile and can integrate acquisitions or new locations much quicker than before. This agility enables us to reach out to the business better. But what’s just as impressive is the impact on the cost side of things: the new setup lets us make significant savings.
Can you be a little more specific about the costs?
Schmied: That’s actually very difficult, because the cost savings in the infrastructure area can’t only be attributed to the cloud strategy. A large part of the savings result from the increased use of internet technology in the WAN area. By switching to a hybrid WAN, we were able to radically cut our MPLS costs many years ago. Plus, we completely dismantled two large data centers. The monthly costs were considerable for just the required network infrastructure.
Calculating the precise costs is a complex exercise: you have building costs, hardware costs, internal and external operating and maintenance overheads, personnel costs etc. It’s very difficult to work out what exactly is being saved where. Today I’m paying less here, but a bit more there. What ultimately counts is how much it costs overall. And today that figure is significantly lower, when all’s said and done. With the size of our operations, we’re talking about several million francs.
"We protect our data and applications with the same high standard – regardless of whether the infrastructure is running in the public cloud, the private cloud or at certain locations still on our own hardware."
Urs Schmied, Head IT Infrastructure
How do you deal with the issue of data security?
Schmied: Of course, the most important aspect here is where the data is stored. With SaaS that’s a bit complicated and it isn’t always entirely transparent. With IaaS, on the other hand, we always know where the data is stored and so we’re also familiar with the legal framework. In concrete terms, we operate IaaS services at four locations: in Frankfurt, Chicago, Singapore, and Australia.
Where would you say the challenges lie in pursuing a cloud-first strategy?
Schmied: The biggest danger with the cloud is definitely that you lose control as a company. You relinquish control of the use of the services you get from the cloud. But you also lose control over the running of the infrastructure, for instance the implementation of global guidelines.
Let’s first talk about the control of the use of the services...
Schmied: When you lose control, that can very quickly become very expensive. «Pay-per-use» is certainly very convenient, but it can also be deceptive. Here’s an example: let’s assume a business needs computing power and sets up an instance to run as an IaaS. That’s straightforward and very quick to do. As long as the resources are actively needed, it’s a very efficient thing. But if these resources are no longer needed and are not terminated, the costs continue to accrue without generating a tangible business value. So it’s very important to make transparent to the business on a regular basis which services in the cloud are currently being used.
Can you describe this process in a bit more detail?
Schmied: By sending a showback to the application owner, we provide transparency for the business. We submit a monthly report outlining what IaaS costs are incurred per application. So those responsible have complete control over how much computing power they use for their application. The good thing about this is that resources that are not being used are immediately identified and shut down. In fact the discussion goes even further: we’ve now taken to switching off certain applications over the weekend, because they’re not needed. To start with I found this idea a bit strange, but once you begin doing your sums, you can make huge savings. And then you very quickly have the support of the business.
That sounds like close cooperation with the business...
Schmied: Exactly, and we want to further expand this cooperation in the future. The business is happy to take on responsibility because it can directly control the positive effect – performance on demand with lower costs. Our goal is to offer the business a self-service setup by providing an appropriate degree of automation. This puts the application owners in a position to manage their virtual servers themselves.
You also mentioned the danger of losing control over the running of the infrastructure...
Schmied: Yes, that’s a very important issue. It is crucial to reduce the operational complexity of a hybrid infrastructure environment to an absolute minimum. How do you guarantee that all security guidelines are followed? How do you guarantee compliance? How do you guarantee the availability and performance of the applications, regardless of whether they’re running on SaaS, PaaS, IaaS or on-prem?
And how do you do that?
Schmied: We delegate this task to Open Systems – the company is in a position to offer both the network layer and the application and security layer from a single source. That can’t be taken for granted, and it helps us no end. After all, I don’t want to have to call seventeen different people when there’s a problem somewhere. With the help of the 24x7 operations we can get to the bottom of any problems right away. Open Systems is responsible for the running and monitoring of the entire global network in the Sulzer Group, which currently adds up to 187 locations around the world. This collaboration has been in place for over 20 years and works very well. Specialized partners are an important part of our operating concept. The openness and transparency of collaboration is very productive for both sides. Open Systems is a partner we can really talk to. Not just when there’s a problem somewhere, but also about new requirements such as our cloud strategy. The engineers were likewise involved from the start and had to do quite a bit of development work to virtualize the services so that they would also run in the cloud.
Does this mean you have full control over the entire hybrid infrastructure environment?
Schmied: Correct, and all with a single point of contact. We protect our data and applications with the same high standard – regardless of whether the infrastructure is running in the public cloud, the private cloud or at certain locations still on our own hardware. Everything is highly standardized, based on a ruleset, monitored with clearly defined processes and visualized in the Mission Control Portal. That gives us security, stability and transparency.
What advice would you give to a colleague who wants to take the plunge into the cloud?
Schmied: A lot of companies have a healthy fear of the cloud. And I must admit, the first step does indeed take some courage. But the advantages are so great that you don’t really have any other option...
"Everything is highly standardized, based on a ruleset, monitored with clearly defined processes and visualized in the Mission Control Portal. That gives us security, stability and transparency."
Urs Schmied, Head IT Infrastructure