Coronavirus: the IT infrastructures have to adapt their sizing

Coronavirus: the IT infrastructures have to adapt their sizing

2020-03-14 -

How does the Coronavirus-related emergency affect the sizing of IT resources and how important is it to upgrade infrastructure?

With the Coronavirus the limits of infrastructures, as we have used them to date, are manifested in all their complexity.

In some cases, these are excess borders, in others of limitations that cause a shortage in services. For example, the airports proved to be over-structured with respect to the request, while hospital infrastructures and, in fact, IT infrastructures, are under-structured with respect to demand.

Covid-19 is profoundly changing the habits of the population and services will adapt to a new way of working and relating: nothing will be the same as before.

Sizing is a central issue for the efficiency of IT infrastructures, and from now on it will be necessary to rethink the ‘capacity planning’ considering the new parameters that emerge in this critical and unusual circumstance.

What does sizing mean?

Sizing is the set of essential resources to be able to provide services to companies and the population at its best. In the case of IT, sizing is defined by the volatile memory in which the data is stored (Ram), and by the amount of central processor units (Cpu).

In the field of e-commerce, you are used to thinking about the number of contemporary customers that a marketing campaign can reach and, therefore, to something predictable.

If, on the contrary, you imagine the functioning of an institutional site that publishes news of interest to millions of people, the architecture of the same, calibrated for a few tens of thousands of users, will have to be rethought.

What parameters to consider for efficient architecture and adequate sizing?

  • Number of services to manage

Once the census of services has been carried out for each site, the following parameters must be estimated:

  • Number of concurrent users for the single service
  • Average application response times of the service

While application response times can be measured, the number of concurrent users is more difficult to estimate, because it mainly depends on the service.

Two examples:

  • e-Commerce: normally, the peak of data traffic is recorded on a promotional campaign via newsletter. The calculation of the number of users who connect in the same time frame is equal to about 30% of the number of emails sent (percentage that depends on the quality of the original database of the contacts).

So, if the newsletter is sent to 300,000 emails, in a shopping period, there will be about 75,000 simultaneous users.

  • Site of public interest: take for example, a public entity, if the published news is useful for the whole population, the calculation must be made considering the population able to use the service.

The average digital literacy percentage in Italy is 70%, and despite considering the high internet access capacity provided by the current mobile, not all users will connect simultaneously to the same site .

In practice, at some point in time, some will be engaged on a particular site, while others will be engaged in other activities.

We define ‘simultaneous users‘ as those who connect to the same domain within 30 minutes, while “simultaneous users” will be those who connect to the same site within 1 minute. Based on our experience, concurrent users represent 30% of the digitally literate population.

Translating into numbers, and taking for example a region that has a population of 5,000,000 people:

Population interested in the message: 5,000,000.00

Dig. literacy 70%: 3,500,000.00

Users available 30%: 1,500,000.00 concurrent users

Concurrent users: 35.000,00 concurrent users

In the current situation determined by Covid-19, the public interest sites are undergoing a real assault. On the basis of the indications described, and in light of the complexity of the moment, it is reasonable to assume that the number of simultaneous users can reach 100%, quantifiable twice as much compared to a normal situation. Going back to the example above, the number of contemporary users is estimated at 70,000.00.

As we have seen, the sizing of IT infrastructures becomes a key factor in ensuring operational continuity and reliability in the provision of services. Functional and efficient infrastructures, able to respond to new application and service needs, are the starting point to face any type of challenge, current and future.

Oplon ADC ensures scalability in vertical and parallel load balancing architectures.