You are here

Cloud Services

Interview with Predica on Cloud Computing & Security

Clutch spoke with Tomasz Onyszko, the CTO of Predica, about cloud migration, storage, and security. Tomasz explained the differences between cloud models as well as IaaS, PaaS, and SaaS cloud computing

Learn more about Predica on their Clutch profile or at predica.pl.

 

Background

Please describe your organization and your position. 

I am the CTO of Predica, a services company working with the Microsoft technology stack. We were established eight years ago, and currently have around 100 consultants working in 20 different countries across the world delivering on-premises and cloud-based projects for customers.  

What services do you offer relating to the cloud?

The services we deliver always come down to finding a solution to the customer’s problem and providing value. We are turning to cloud solutions by either moving machines to the cloud in simpler cases or building applications and solutions on cloud stack. This can include Software as a Service such as Office 365 or new applications built on Platform as a Service technologies.

Challenge

What has been the biggest challenge you've witnessed that organizations face as they migrate to cloud computing?

It’s easy for a company to jump on board and say that they want to migrate to the cloud, but the first question they need to ask is, “Why are we moving to the cloud?” This question will answer what potential benefits are given by the environment. The biggest challenge for us, as consultants, is to drive this discussion, and build awareness around what the cloud can do, and how it can be used to drive business value. There is a lot of buzz around this topic right now, but there are different flavors, providers, and technologies available. 

Solution

What types of companies are the best fit for Public, Private or Hybrid cloud? What are the pros and cons of each?

In general, the cloud model steers clearly away from the classic IT model, where companies needed to consider which server was running which application, or what type of hardware was needed. A cloud model of any type enables businesses to get the right resources when they need them. Companies can start to move much faster and enable different business models for themselves. This is why people want to go to the cloud. 

No one company is a perfect fit for a certain type of cloud. Some companies are restricted to private clouds, possibly through local regulations. For example, finance business in Poland has this restriction imposed by the local regulatory body. These restrictions can influence what type of cloud is adopted. Companies opting for a private cloud either have to deal with local regulations or have a degree of internal inertia. 

In the end, companies which can allow themselves to take the most flexible approach are going for the public cloud model. But the reality is that most companies now are using a mix of the two – they still have their own datacenters, but are using public infrastructure at the same time. Companies that use the hybrid cloud models see the benefits and speed of public solutions, but also need to keep hardware on-premises. Hybrid is a very broad term and can encompass a number of models.

What types of companies are the best fit for IaaS, PaaS, and SaaS?

The cloud can be described as renting computing power and storage. Infrastructure as a Service is a simple first model adopted by the cloud. Typically, when a company has its own infrastructure, such as a data center with servers, they need to maintain it. If they buy a new server, they also need a new rack, and to connect the server to a power supply. This can be costly and slow compared to IaaS which allows companies to rent that infrastructure using computers hosted by a provider. When using IaaS, the complexity of buying, hosting, and maintaining their own hardware is removed. Instead of dealing with actual machines, the user requests an amount of computing power and network. This is the most basic level of cloud and is typically used in a lift-and-shift model – moving existing computers to the cloud in order to get rid of datacenters, or moving applications and SQL databases written for local servers. 

Following this, someone scratched their head and decided to increase the level of abstraction – if we need a SQL server, why do we need to deploy it as a computer? We just need the service. This is how Platform as a Service works – instead of focusing on the boxes running a task, there is a set of services such as web servers or SQL servers which provide building blocks. For example, when leasing a PaaS for analytics gathering, we are not thinking about how the infrastructure for it will be built; we’re simply uploading the data to a service that runs those analytics. We don’t have to think of the infrastructure underneath since there are API interfaces for it. We can build business applications just connecting those services.

At Predica, we use a set of PaaS implementations on Microsoft Azure, including cognitive services which can recognize a voice or a picture. We can upload audio files and receive transcriptions back within a few minutes, without worrying about how the process occurred. We are also working on a project for a company manufacturing water tanks which require monitoring. We have created a solution using only the available building blocks from the PaaS. PaaS is targeted on building complex applications rapidly, in cases where the business doesn’t want to handle the infrastructure, and only wants to build a solution. 

Software-as-a-Service is a ready solution which doesn’t require connecting anything together; it can be simply switched on. As an example, if someone needs a CRM system, they can simply start a new Salesforce instance. If they need email, they can start a new instance of Office 365. 

If our client companies need a service, their first step will be to look for a SaaS option. With this, there is no requirement for a specific fit – any company can use the service, if they need it. In the past, if someone wanted to deploy a CRM, they would have to go through a 6-month project. Right now, they can do it in five minutes, test the scenario for a month, and, if they don’t like it, scratch the implementation and start a new CRM. The solution is one credit card away.

In the end, companies which can allow themselves to take the most flexible approach are going for the public cloud model. But the reality is that most companies now are using a mix of the two – they still have their own datacenters, but are using public infrastructure at the same time. Companies that use the hybrid cloud models see the benefits and speed of public solutions, but also need to keep hardware on-premises. Hybrid is a very broad term and can encompass a number of models.

Features

What are some of the factors that companies need to consider when selecting a storage platform?

Cloud storage is rarely purchased as a separate service. Companies purchase it together with other cloud components. Productivity storage offerings like Box, Dropbox, or Google Drive are purchased with a larger productivity suite. For example, when buying Office 365 for a client, we will choose OneDrive.

In terms of storage for large quantities of data, companies will typically onboard to a cloud provider like Amazon Web Services, Azure or Google. We always look at storage from the wider perspective of which computing provider or productivity suite was chosen. 

It is always important to consider security, the ability to control access, and compliance requirements when choosing a storage provider. 

What are some of the factors that companies need to consider when selecting a storage platform?

The most important thing that companies should realize is that the old way of conceptualizing security is over. Some people still believe that they are protected when they are within their own boundaries or on their own network. But this thinking is outdated. 

Security is a very broad term, encompassing not just a single server, but access to data in general. Security departments need to consider how people are accessing information, where they are located, how we can protect the perimeter, and what the perimeter actually is. Ten years ago, a perimeter was the local network, but now, users can access information anywhere, anytime, using any kind of device. The challenge is how we can provide protection in this context. 

When companies are thinking about onboarding cloud technologies, they have to consider the broad spectrum of people working on different networks and devices. After moving to the cloud, the main challenge comes from the lack of understanding from security departments around what protection solutions are available, and how they differ from on-premises infrastructure. 

Within both enterprises and smaller companies, the security departments often have a good understanding of how to protect local networks, but little understanding of what the behavior is in the cloud, and how to apply security and protection, accordingly. My advice to companies adopting the cloud is to approach security from a user perspective. Think about the different resources users have access to, and how those can be protected. Companies need to focus on identity management.

When a company deploys a new cloud solution, they may immediately begin granting privileges. While cloud providers may know how to protect their data centers, companies themselves need to focus on how to protect the data and access to those services. This requires an understanding of how encryption works, how and where data is stored, compliances from a load perspective, who has access to it, and who manages that access. These are the most important factors of managing security on these platforms. Each platform has its own tools and mechanisms for security management.

In order to be more secure, we need to have an understanding of changes in the security model and the paradigm shift.