Sergey Migalin: ‘The Volga Region is a big region where demand for services of data centres is growing’
The vice head of Rosenergoatom about prospects of Data Processing Centre in the Tatarstan IT town, eco-friendly use and state of affairs in the data processing market
Why did Rosenergoatom choose Tatarstan as a site to build a Data Processing Centre (DPC) and what characteristic and parameters will Innopolis DPC will have? Vice Director General and Director of Economy and Finance at Rosenergoatom Sergey Migalin answered these questions in an interview with Realnoe Vremya. According to the speaker, the centre created in the republic will allow meeting growing demand for services of commercial data centres in such a big industrial region as Volga. Citizens of the smart town whose houses can be heated with the heat generated by the DPC will also feel the benefit of the new neighbour. Read in our talk about where else similar data processing centres can appear and what the general situation in the DPC market in Russia is like in general.
“We need a DPC in the Volga region to support the corporate cloud”
Sergey, we’d like to understand why precisely Innopolis was chosen as a site for the DPC. For instance, there is no nuclear power plant as an uninterrupted and powerful source for energy saving, while the territory is far from operators’ broad communication lines.
The logic of our choice had three key moments. Firstly, it is the geographical location. We need a DPC in the Volga Region to support the corporate cloud, also to store and process data from our nuclear power plants. So the Volga Region is a big industrial region where demand for services of commercial data centres is growing, and here we enter the market with a ready proposal. Previously, we considered the possibility of building a data centre near the Balakov Nuclear Power Plant, but as the infrastructure in the area of the nuclear plant was already created, we decided to find another site. Though, from a perspective of electrical energy, a DPC near a nuclear power plant is convenient, without doubt.
We need a DPC in the Volga Region to support the corporate cloud, also to store and process data from our nuclear power plants
Secondly, it is electrical energy you’ve already mentioned. We addressed this issue with the management of Innopolis SEZ and received an offer of optimal tariffs to connect electrical energy. Obviously, this isn’t the prime cost of the nuclear plant, nevertheless, it is a stable high-voltage connection. Choosing among the parcels the SEZ offered, we chose the one that will be optimal in terms of technical hook-up.
Thirdly, and most importantly, after we began to consider the sites, we understood that the team Innopolis is a team that really wants to achieve a result. We are fully convinced that it is crucial for such projects. If there are a mutual desire, constructive outlook and drive, one will succeed. If you have a super site but you don’t have a team of single-minded people, nothing will work. This probably became the decisive argument in favour of Innopolis.
I also want to note that our cooperation with Innopolis won’t limit only to the construction of the data centre. Our subsidiary CONSIST-OS has an office operating in Innopolis since 2019 where the key centre for development, support and modernisation of robotic software for both Rosatom’s Electrical Energy Division and external customers.
Moreover, we plan to create a laboratory to develop cloud services and a competence centre for equipment predictive analytics in Innopolis DPC. We are also interested in joint work with Innopolis University both as an area to train staff for our projects and provide University with access to our scientific cloud in Kalinin DPC.
We plan to create a laboratory to develop cloud services and a competence centre for equipment predictive analytics in DPC Innopolis. We are also interested in joint work with Innopolis University
“Several billions of rubles will have to be invested in server infrastructure and development of the commercial cloud”
What are the technical parameters of the DPC in Innopolis?
It will be quite a big DPC: its target consumption is 16 MW, which is quite much. To compare, consumption of Russia’s biggest Kalinin DPC in Udomlya is 48 MW. We hope to reach a thousand racks in the short run, but we should understand that such calculations are relative enough, as the DPC used to be traditionally calculated as 5-6 kW racks, while the situation is different now. The amount of the first stage of the project in Innopolis is 500 racks.
From a technical perspective, it is a fixed facility consisting of two buildings with an administrative sector for a small office to service and support clients apart from the DPC. Now it is not enough to simply place equipment for commercial DPCs — clients need a number of additional services: warehouses modules, co-working space, etc. We plan to design Innopolis DPC as a fully functional complex.
Engineering components account for main construction costs. First of all, it is cooling, fire fighting systems and provision of uninterrupted electrical supply with the use of diesel. In the current project, we plan to spend about 3 billion rubles to equip the DPC.
About a third of the DPC will be filled with our own equipment where will unfold Rosatom’s commercial cloud. At the same time, our protected corporate segment will certainly be located there. Several billions of rubles will have to be invested in server infrastructure and development of the commercial cloud. There aren’t more precise evaluations here now, as the commercial cloud is a scalable and dynamic story.
By late 2022, the DPC will start functioning as a fully-fledged infrastructural facility for the whole Volga Region
From the point of view of the project’s deadlines, we’ve already launched the design and plan to get final design solutions by the spring of 2021. We can start the construction of the first phase by the middle of the next year we plan to launch in late 2022. By late 2022, the DPC will start functioning as a fully-fledged infrastructural facility for the whole Volga Region.
Last year we talked with the management of Innopolis SEZ, and we were told that the project would likely be implemented on two sites: one is designed for modular DPCs. Is this information confirmed?
In general this isn’t considered seriously now. We think that it isn’t efficient to build such DPCs in Innopolis. The construction of quickly erected modular DPCs is costly. Their cost is high enough, and, as a rule, the use of such DPCs is feasible only where a high speed of construction of infrastructure is important. The modular DPC is also optimal in complicated conditions, on hard-to-reach territories where a full cycle of construction works is impossible and the speed and minimal expenses on local contractors is important as well.
There is fierce price competition in the market of DPCs. It is no surprise that a lot of operators consciously, and maybe forcedly, dump by offering prices that don’t comply even with the prime cost of maintenance of fully-fledged infrastructure. It is hard for a customer to distinguish quality.
From a perspective of the price for services, the DPC significantly differs. If we take the regional market, the gap will be considerable — from 20,000 rubles for an average rack and even less, while it is a very low level that doesn’t cover real costs on engineering infrastructure) to a reasonable bar of 40-50,000 rubles.
The construction of quickly erected modular DPCs is costly. Their cost is high enough, and, as a rule, the use of such DPCs is feasible only where a high speed of construction of infrastructure is important. The modular DPC is also optimal in complicated conditions, on hard-to-reach territories where a full cycle of construction works is impossible and the speed and minimal expenses on local contractors is important as well
“We are ready to integrate this team to our project, for instance, as a “sales window”
How much is the forecasted revenue of this DPC if we consider the experience of the other sites? For instance, it became possible to earn half a billion rubles in Kalinin in 2019.
Of course, we have a financial and economic model and certain outlooks, but I won’t comment on this yet. When you have several DPCs and there is the infrastructure that’s geographically distributed, first of all, the cloud where you sell cloud services, it is hard enough to determine they are sold from Innopolis or Moscow. It is an absolutely abstract thing. This is the essence of our new virtualised world.
From this point of view, on the one hand, the Innopolis DPC is an infrastructural hub in our geographically distributed system. On the other hand, it is not a separate site. In general in our programme of the geographically distributed network of DPCs we plan to exceed 15 billion rubles in five years.
Again, in the talk with representatives of the SEZ we learnt that, to put it roughly, Innopolis will provide a site, Rosenergoatom — experience and investments, while Kazan IT Park can take on the management of the DPC. They explained that their site has its own DPC that had been operating for a while. Consequently, they have experience in managing such facilities.
It is true, IT Park is a team that has considerable competences directly in Tatarstan. We focus on partnership and are ready to integrate this team to our project, for instance, as a marketing agent, “sales window” on the territory of Tatarstan and other neighbouring regions.
When you have several DPCs and there is infrastructure that’s geographically distributed, first of cloud where you sell cloud services, it is hard enough to determine they are sold from Innopolis or Moscow
“Environmentally, this heat has no harm”
DPCs generate a lot of heat, and thermal energy often dissipates in the air. However, some countries learnt how to use it effectively — data centres are connected to the system of centralised heating. Do you plan to introduce such a system in this facility?
Environmentally, this heat has no harm. Economically, of course, it would be useful to use this heat somehow. But we don’t have such plans in the short term. As I already said, Innopolis DPC is built with a commercial purpose.
At this moment, developed, mass systems that would allow getting rid of heat economically to use it for something useful simply don’t exist. One can’t take a solution out of the box and immediately use it. Long and expensive research in this area is required. We already have an idea of creating an R&D DPC where we will test new technology, but this certainly won’t be done in the next 2-3 years.
We should also understand that we can’t be directly compared with Europe. Firstly, our energy system is different. Secondly, our tariff fixing is completely different. The cost of electrical energy in Europe is about 3-4 times higher. The profitability of such solutions when you have a threefold margin in the retail price for a client is high, of course. In our case, such stories don’t pay back at the moment.
It is a pity, it would be interesting if this small town would be heated this way.
I think everything is possible. If we see a big growth of the market and, first of all, its civilisation and transformation, the profitability of commercial projects of DPCs will allow developing related areas. From this point of view, indeed, there are interesting ideas. For instance, an open swimming pool heated with the heat from the DPC could be created on the roof in the centre of Innopolis.
The cost of electrical energy in Europe is about 3-4 times higher. The profitability of such solutions when you have a threefold margin in the retail price for a client is high, of course. In our case, such stories don’t pay back at the moment
We will come to it, without doubt. The question is how fast, what money with and who will be ready to invest. Now we have a stage of fierce competition, as the marketing is getting strong, the number of both DPCs and players are suddenly increasing, at the same time, we are switching from a model of cheap and not very good service to hyper DPCs.
The DPC in Innopolis is just a part of Rosatom’s global chain of data centres. Could you tell us about other projects as well?
Our global strategy looks this way: we will certainly build big DPCs or will purchase them with the possibility of modernisation in Moscow, Saint Petersburg, Innopolis. So we will get an absolutely clear link between Moscow, the Northwest and the Volga Region. We separately considered options to cover Ural Federal Okrug but at the moment we don’t see high demand for services of data centres there.
It is also possible that we will build a data centre near the Kola Nuclear Power Plant if this is in demand within the development of infrastructure of the Northern Sea Route and logistics along the Arctic. But this is also a prospect of the next 3-5 years, not earlier. We don’t plan to build additional powerful data centres in other regions yet.
We also use an approach in which the company focuses not only on its own infrastructure but also a collaboration with other market players. We are focused on partnership, as the market of DPCs and clouds isn’t simply dynamics, it must know how to change parameters to meet clients’ requests, both territorial and in terms of service. So the exchange of capacities is a completely normal practice in the market.
For instance, MTS is building a DPC in partnership with us on a parcel near Novosibirsk Chemical Concentrates Plant — it is an enterprise of Rosatom that produces nuclear fuel. We also consider joint projects with other companies we will soon be ready to tell you about.