Hybrid cloud: what is it and what is it for

Hybrid Cloud is an infrastructure that combines the public and private clouds to deliver the best of both worlds. Find out everything you need to know about it.

The public cloud has many benefits for businesses. However, it also brings its share of constraints and challenges. In fact, many organizations prefer to stick to the private cloud or an on-site infrastructure.

However, a study published by Nutanix in 2018 reveals that 91% of IT managers of companies consider that the ideal model is that of the hybrid cloud , which combines the benefits of the public cloud with those of the private cloud .

Similarly, by 2019, 451 Research analysts estimate that 69% of businesses will rely on hybrid cloud environments and 60% of workloads will run on some form of cloud service. So, transitory strategy or optimal compromise? Find out what you need to know about the hybrid cloud.

Hybrid cloud: definition

Before defining the hybrid cloud, it is important to understand the distinction between public and private clouds . In the case of a private cloud, the servers are dedicated to a single company. These servers can be on-site, or off-site.

In the case of a public cloud , the servers are shared between the different customers of a provider. The servers are still off-site, since they are located in the provider’s Data Centers.

According to Forrester Research, the hybrid cloud is about connecting one or more public clouds to a private cloud or traditional on-premise data center infrastructure. To put it simply, it is a mix of on-site and off-site IT resources.

More elaborately, the hybrid cloud is a cloud environment consisting of on-premise private cloud resources combined with third-party public cloud resources connected to each other by an orchestration system .

According to the “official” definition of the National Institute of Standards and Technology , the hybrid cloud is “a cloud infrastructure consisting of two or more separate cloud infrastructures that may be private or public and which remain unique entities but are connected by a single technology. standard or proprietary allowing the portability of data and applications “.

Hybrid Cloud: What are the benefits?

One of the key benefits of the Hybrid Cloud is that it can move workloads and data between the public and private clouds in a flexible way based on needs, demand, and costs. As a result, organizations benefit from increased flexibility and additional options for data deployment and usage.

The flexibility is indeed a major advantage of the hybrid cloud . For good reason, in the case of an on-site infrastructure, resource management requires time and money. Adding capacity therefore requires upstream planning. On the contrary, the public cloud is already ready and resources can be added instantly to meet the needs of the business.

Thus, by leveraging the hybrid cloud, a company can leverage public cloud resources when its needs exceed the resources available on private clouds, for example during seasonal peaks. The hybrid cloud therefore provides the elasticity needed to cope with demand variations that can be linked to multiple factors.

In parallel, as in the case of on-premise private cloud storage, the hybrid cloud provides quick access to the most critical data . It is therefore possible to keep frequently used data on site, and to transfer “cold” data to the cloud.

In addition, the hybrid cloud helps reduce business costs with the low cost of IT resources available on the public cloud. Indeed, most public cloud providers offer their customers to pay only for the resources they consume. Unnecessary expenses are avoided.

The hybrid cloud is also ideal for processing Big Data . For example, it is possible for a company to use hybrid cloud storage to store its data and to perform analytic queries on the public cloud where Hadoop (or other) clusters can be scaled to fit the computing tasks. more demanding.

Hybrid cloud: what are the disadvantages?

Despite its many advantages, the hybrid cloud is not suitable for all situations. For example, for small businesses with limited IT budgets , the costs of installing and maintaining private servers may be too high. So it’s best to stick to the public cloud.

In addition, an application that requires minimal latency is not always suitable for the hybrid cloud . It may be better to opt for an on-site infrastructure. However, it all depends on the chosen cloud provider. It is therefore important to consider both the application and the resources offered by different cloud providers to decide whether to opt for the public, private or hybrid cloud.

Hybrid cloud: who are the main suppliers?

Many companies today offer hybrid cloud infrastructures. Among the market leaders, there is of course Amazon Web Services also known as AWS . The US company offers storage, networking, security, application deployment and hybrid cloud management tools.

For its part, Google Cloud has partnered with Cisco to offer an open hybrid cloud architecture . This architecture is based on the Istio open source containerization service, which allows the user to connect and manage connections between on-premises and cloud services.

Another supplier is HPE, which offers a customizable solution based on the needs of the company. Similarly, Dell EMC Hybrid Cloud is a solution that enables organizations working with large workloads to perform their digital transformation for greater efficiency.

Recently bought by IBM for $ 34 billion , Red Hat offers a variety of hybrid cloud products like RedHt Openshift and the Red Hat Cloud suite. Other vendors include Oracle, VMWare, SAP with its HANA platform, and Citrix .

RaaS and Cloud Robotics: When robotics meets the Cloud

RaaS or Robotics as a Service and Cloud Robotics are two new models that combine robotics with Cloud Computing. Find out everything you need to know about it.

Because of the many benefits they bring, cloud services are becoming increasingly popular with businesses in all industries. In fact, service providers respond to demand by offering an ever more diversified offer.

The first categories of cloud services, and the most used to date, are SaaS (software as a service) , PaaS (platforms as a service) and IaaS (infrastructure as a service). However, to support the digital transformation of businesses, cloud services are offered for new technologies . This is for example the case of automation, virtual reality, artificial intelligence, but also robotics.

RaaS or Robotics as a Cloud Service: Definition

Robotics as a Service (RaaS) is a model that combines cloud computing, artificial intelligence, robotics, and shared services . This template allows users to store data collected by robots on the cloud.

For example, robots located in different locations such as warehouses or stores are connected to the internet and the data they collect is transmitted directly to a centralized cloud storage system . This makes it easy and quick to access and process data. RaaS is particularly useful for companies that manage fleets of several hundred robots performing different tasks.

The RaaS concept also integrates robot rental services . These services allow companies to rent robots and subscribe to a cloud service rather than having to buy such machines. This allows users to save money, but also to gain flexibility.

Cloud Robotics: what is it?

Parallel to RaaS, another concept marrying the Cloud to robotics is Cloud Robotics . It is about using the resources of Cloud Computing to increase the memory, computing power, collective learning and interconnectivity of robotic applications. Thus, when computing or storage power requirements exceed the capacity of the robot, it is possible to exploit the massive resources of a data center via the cloud.

In addition, the cloud also speeds up the learning of robots . For example, if a robot needs 150 hours to learn a task using its own artificial intelligence resources, the collaboration between 150 robots can allow the AI ​​to learn how to perform the different parts of a task in hardly an hour. Robots can indeed share their knowledge via the cloud. This interconnection also allows them to better coordinate.

The connection to the Cloud also allows the robot to access an extensive library of objects to learn to identify the elements that make up its environment. Object recognition can then enable the robot to more efficiently perform tasks such as sorting, cleaning, or using devices.

In addition, the exploitation of Cloud resources makes it possible to involve robots by reducing their computing power and the associated cooling power systems . Robots can be lighter, cheaper, and more autonomous.

In general, Cloud Robotics is used for tasks that do not need to be executed in real time . Thus, local resources can be preserved for more urgent tasks.

RaaS and Cloud Robotics: some examples

Among the various suppliers of RaaS, there are companies like Aethon, Locus Robotics, Savioke, Sarcos Robotics, or Redwood Robotics . The latter offers its customers a Cloud solution and preprogrammed rental robots to enable them to enjoy the benefits of robotics more easily, cheaply, and flexibly.

In the coming years, RaaS services will multiply. By 2026, ABI Reserach estimates that the RaaS market will reach a value of 34 billion dollars against 217 million dollars in 2016.

With regard to Cloud Robotics, there are several examples. For example, Google’s autonomous cars are a type of cloud-connected robot . They access Google Maps data and images stored on the cloud to recognize their environment. In addition, they collect traffic and traffic information and transmit it to the cloud. Note that Google Cloud will launch its platform dedicated to robotics in 2019 .

further example is that of Romo : a small robot connected to the Cloud to $ 150. The device uses the computing power, the camera, the microphone and the speakers of the iPhone to which it is connected. Thanks to the Cloud, the robot is able to better recognize its environment. It can be used to play games, or for telepresence purposes.

Cloud Management Platform: what is it and what is it for?

A Cloud Management Platform is a platform for easily managing one or more cloud infrastructures from an intuitive and unified interface. Find out everything you need to know about it.

The public cloud has many advantages to offer companies : flexibility, scalability, cost reduction … however, without carefully thought-out strategy, without clear objectives , it is impossible to take full advantage of these strengths. Companies that do not prepare for migration to the cloud often end up returning to their traditional data centers or the private cloud because the public cloud is ultimately a waste of time or money for them.

The management of Cloud applications often involves repetitive tasks, and requires knowledge, skills for which it is still difficult to find experts. This is particularly the case for large environments. That’s why Cloud Management Platform (CMP) or Cloud Management Platforms are becoming increasingly popular.

Cloud Management Platform: What is it?

A Cloud Management Platform is a suite of integrated software tools that enable a company to monitor and manage public, private, hybrid, or multi-cloud public cloud infrastructures.

There are different types of CMPs, and the features offered vary from platform to platform . Some provide more tools than others, and some are specifically dedicated to specific industries. In addition, cloud management platforms can be provided as on-premise applications, but some vendors offer them as cloud SaaS services .

Cloud Management Platform: What are the different types?

The most basic Cloud Management Platforms can only manage one facet of the cloud infrastructure. For example, Evident.io offers only security features for the cloud, while Skeddly allows only automation and cost optimization for Microsoft Azure. However, the best recent CMPs provide both solutions for cost management, security, and cloud automation.

The management platforms “multi-cloud”  to manage multiple cloud infrastructures from different suppliers from a unified interface. For example, it is possible to manage AWS, Azure, Google Cloud Platform and IBM environments from the same platform.

Finally, hybrid cloud management platforms can manage a cloud environment and an on-premises environment from the same interface. These platforms are therefore ideal for companies that opt ​​for a hybrid cloud strategy.

Cloud Management Platform: what are the features?

Each cloud management platform has different features, but the Cloud Standards Customer Council believes that CMPs must offer certain features.

They must first offer a self-service interface , and reporting and analysis capabilities to provide advice on the company’s cloud consumption. They must also be able to monitor and manage cloud services to enable planning and deployment of workloads.

CMPs are also supposed to help manage cloud resources : virtual machines, storage, networks … and ideally allow the migration of resources between different environments.

With regard to financial management, the CMPs must be able to track and distribute the company’s expenses in cloud computing , generate reports and predict future costs. Finally, these platforms must allow the administrator to set rules to control the use of cloud resources and provide security features such as encryption and access management.

What are the best CMPs on the market?

There are a large number of cloud management platforms. Among the most popular are Wrike, IBM Cloud Orchestrator, Apache CloudStack, Sympantec Web and Cloud Security, ManageEngine Applications Manager, OpenStack, AppFormix, ServiceNow Cloud Management, Centrify Application Services, Cloud Lifecycle Management, MultCloud, Bitiumn Zoolz Intelligent Cloud, RightScale Cloud Management, Microsoft Azure Cost Management, CloudHealth, Cloudcraft Morpheus, Cloudify, xStream, vRealize, Cisco CloudCenter, Sclar, Red Hat CloudForms, and HP OneSphere.

Cloud Native: what is it and what is it for?

A Cloud Native approach is to develop an application taking advantage of the various benefits offered by the Cloud.Find out everything you need to know about it.

According to the Cloud Native Computing Foundation (CNCF) , founded in 2015 by the Linux Foundation, Cloud Native technologies enable organizations to create and run scalable applications in modern and dynamic environments such as the public, private or private cloud. hybrid .

Examples of technologies include containers , the mesh service, microservices, immutable infrastructures, continuous integration or declarative APIs. Applications created with this approach are typically based on an infrastructure as an IaaS service or a platform as a PaaS service that will then be combined with the previously mentioned tools.

What is the Native Cloud for?

A Cloud Native approach has several advantages. First, it allows companies to transform their application ideas more quickly into products available on the market . It not only speeds up change, but also reduces risk.

In addition, Cloud Native technologies also enable increased scalability for applications . As companies attract new users to more regions and more devices, it is possible to maintain application responsiveness and keep costs down.

Another major benefit is that it usually allows businesses to spend less money on accommodation. Thus, it is possible to effectively reduce the loads .

What are the differences with on-site applications

Cloud Native applications require a very different architecture than traditional enterprise applications , which is why they have many differences with on-premises applications.

First, while on-premise applications designed to be run on corporate servers are usually written in traditional languages ​​(C / C ++, C # …), Native Cloud applications are written in web-centric languages like HTML, CSS, Java, JavaScript, .net, Go, Node.js, PHP, Python or Ruby.

Cloud Native apps also have the distinction of being always up-to-date and available . On the other hand, on-premise applications require updates during which they are unavailable, and are usually offered with a subscription model by the vendor.

In case of peak use, Cloud Native applications can take advantage of the elasticity offered by the cloud by using more resources. These can then be disabled when usage returns to normal. This is a flexibility that is not available on site.

A source-based application can also be run in a virtualized environment and share resources with other applications. In the event of a failure of one of the Cloud Provider Data Centers, the stack can be immediately transferred to another region. The risk of downtime is therefore greatly reduced compared to an on-site application.

With Cloud, Cloud Native application management is fully automated with automation and orchestration tools. In addition, Cloud Native applications have a modular design because many of their functions can be broken down into microservices. It is therefore possible to disable certain features or to deploy updates for specific modules rather than for the full application.


IBM revolutionizes machine learning with quantum computers

IBM has created quantum algorithms to perform machine learning on quantum computers to create artificial intelligences much more powerful than those created with conventional computers …

The  “Feature Mapping” is to disassemble the information to access more aspects “purposes” of the data. Currently, Machine Learning already allows you to do this, for example by taking the pixels of an image to place them in a grid according to their color. The algorithms then map the color values ​​non-linearly and break down the data according to their most useful characteristics.

However, thanks to quantum computers , IBM researchers have discovered a way to make Machine Learning significantly more efficient for feature mapping. In a paper published on arXiv, the research team announces having created a “quantum algorithm” allowing quantum computers to perform machine learning in a new scale.

As a reminder, quantum computers take advantage of the strange ability of subatomic particles to exist in more than one state at a time . Thanks to the way these particles behave, operations can be performed faster than on conventional computers, while consuming less energy.

In fact, unlike the bits of classical computers that can only exist in two states (1 or 0), the quantum bits (qubits) of quantum computers can exist in any superposition of these two values ​​and thus store more data. information.

Thus, IBM’s new quantum algorithms make it possible to separate the aspects and characteristics of the data in an even greater degree than with a standard Machine Learning algorithm. In fact, data can be classified more precisely and Machine Learning systems will be more efficient.

IBM takes Machine Learning into the quantum dimension

The goal is to use quantum computers to create new classifiers that can generate more sophisticated data maps. In doing so, researchers will be able to develop more effective artificial intelligences that can for example identify invisible patterns for conventional computers.

For now, IBM says that these new algorithms have not yet surpassed the performance of conventional machines on quantum computers. However, this is mainly related to the fact that quantum computers are still limited by the current hardware constraints.

Indeed, current quantum computers have a computing capacity limited to only two qubits. However, this computing capacity can be simulated on conventional computers. It will therefore be necessary to wait for more efficient quantum computers to emerge for IBM algorithms to achieve “quantum advantage”.

In the meantime, these new algorithms are available in open-source via the Qiskit Aqua library at this address. Developers, researchers and other experts can use it as they see fit.


How Softlayer became IBM’s public cloud provider

Softlayer is a cloud computing company founded in 2005. Its rapid growth in the private cloud market has attracted many companies. Softlayer currently belongs to IBM. Explain.

Softlayer was born in 2005. The expertise of this company is based in Dallas Texas? Deploy Infrastructure in the form of Services or IaaS solution to store data on a large scale. From 2005 to 2013, she quickly developed her architecture. It currently controls 14 data centers and manages 100,000 servers in the United States, Europe and Asia.

His strengths? A solution based on automatic control of physical and virtual devices via API allows customers to deploy public cloud, physical server and private cloud instances via Softlayer VPS access points and VPN.

Brilliant launches for Softlayer

This rapid growth of Softlayer has been noticeably made possible by its acquisition in 2010 most of the stake in GI Partners, a private investment fund. The goal is to get another data center and related service provider, neighbor Theplanet, based in Houston, Texas. The value of the activity: $ 2.35 million.

In 2011, Softlayer expanded its customer base, making it one of the largest private cloud storage companies in the world. The leaders declared at that time hosting 81,000 servers for 26,000 customers in the United States and a turnover of $ 78 million in the first half of this year.

Of course, a company like Softlayer is very interested in big names in Cloud computing, starting with IBM, the company has increased 80% of revenue from this activity in 2012. IT giant sees in This expert IaaS is a way to gain market share in this highly rated field.

It is primarily a way to control the entire value chain around IBM’s cloud services. We found them in Bluemix and the famous Deep Learning solution, Watson. Therefore, Big Blue wants to provide services through its own infrastructure.

Softlayer, Bluemix: IBM chose

In 2013, IBM bought the Softlayer for $ 2 billion. The company keeps its name by becoming a subsidiary of the company. This activity was completed in June 2013 and quickly followed by announcements. It is easier to remember when opening a data center in Paris, especially in Clichy in 2014. The installation of Softlayer technology in France attracted special attention.

However, we must not forget the work done in less than two years by the company. In 2015, Softlayer based in IBM stated that they have 23 data centers in 11 countries around the world. Collaboration with California Data Center manager Digital reality is very important to complete this $ 1.2 billion expansion.

Therefore, Softlayer is at the heart of IBM’s cloud strategy until 2016. But last year saw the decline of traditional Big Blue activities. Cloud revenue has increased 35% over the same period last year, but market share is much lower than Amazon, mastodon has been operating in the industry for more than a decade.

To achieve clarity and clarity from customers and investors, IBM chose to abandon Softlayer brand image. So in October 2016, the company’s executives announced that the name Softlayer would disappear little by little to support the Bluemix brand.

Therefore, it is now possible to provide Softlayer services from IBM Business Accounts. In fact, IaaS expert website always refers to sites that offer Bluemix.

Softlayer and its price

This change of identity comes with Softlayer discounts, meaning that prices are calculated, and this is in the third quarter of 2016.

New price bases can be found on the Softlayer website. The incentives are divided into two entities: physical server and physical server. The first offer starts at $ 0.368 per hour or $ 158 per month for a service in less than half an hour and a storage facility of 500 GB. The second offer for cloud storage is charged from $ 0.038 per hour or $ 25 per month. This time, customers have 250 GB of storage.

This is obviously a facility because the company can provide storage services supported by 56-core processor and 242 GB RAM for very large tasks (Deep learning or data processing in time real).

The market plunged

Although this rate is reduced, this does not prevent large Softlayer / IBM IaaS customers from leaving the ship. This is Whatsapp messaging case that changed ice cream factory and store customer data on the “on-premises” infrastructure of Facebook owners.

Competitors in the Bare Metal server market are dedicated to storing containerized technologies such as Docker Swarm or Rancher platform.

The last thorn at IBM’s feet, the company’s Houston-based infrastructure was no longer active after the storm Harvey attacked the town last weekend. “Innovation or death” is the company’s motto since 2005. IBM will succeed in gaining a position in the cloud? Questions still to explore.