Overview of OpenStack Hybrid Clouds and the Challenges of Building It
The massive introduction of cloud technologies has fundamentally changed the process of deploying new services in the corporate IT infrastructure. In the classical approach, such a process could take several days, weeks or even months - it was necessary to first order a new server, then deliver and install it in the company's data center, manually install the operating system and application software on the server, adjust the application settings and only after that it was possible to run on it a new service. When using cloud technologies, service deployment takes several hours or minutes thanks to the self-service portal, through which the system administrator or service owner can quickly obtain all the resources necessary for the service to work (processors, RAM and disk capacity). In this article we will talk about how to quickly and safely switch to a hybrid cloud based on OpenStack.
There are two main options for switching to cloud computing:
Both approaches have their pros and cons. The advantage of a private cloud is high security (applications and their data do not leave the company), more effective control over the performance and reliability of applications, the absence of dependence on an external supplier and a virtually unlimited choice of virtual machine configurations. The public cloud requires significantly less initial investment in cloud technologies and allows you to pay for resources allocated to applications using the pay-as-you-go model, provides transparency in IT costs, and also significantly exceeds the private cloud in terms of capacity scaling and ease of management.
')
For successful implementation of a cloud infrastructure in a company, it must support both classic vertically scalable high-availability server applications that can work only on a single server, as well as horizontally-scalable applications designed for clouds that can run on multiple servers at the same time and do not require software high availability. The failure of one instance of this type of application does not lead to the failure of the entire service provided by the application.
Since both cloud variants have their pros and cons, the best application for a cloud technology company is a combined approach, in which some applications are deployed in a private cloud (usually business-critical applications and applications with increased security requirements), and all others in the public cloud of an external provider. Such a hybrid cloud usually combines both virtual and physical infrastructure, supports different service delivery models (IaaS, PaaS and SaaS) and different hypervisors (KVM, XenServer, ESXi and Hyper-V).
Hence, the task of ensuring the interaction between applications located in the private and public cloud, requires mechanisms that implement the exchange of data and application templates between these two clouds, and if necessary, transfer between them the applications themselves. Porting applications may be needed, for example, if you need to free up some of the resources of a private cloud due to a sharp increase in the load of one application. Therefore, some applications can be temporarily transferred to the public cloud. In addition, after testing in a public cloud a new version of the code of a business-critical application, it should be moved to a private cloud. Mechanisms are also necessary to allow, when changing the cloud service provider, to transfer all applications from the public cloud of the previous provider to its public cloud. Thus, we need ready-made solutions that guarantee the compatibility between private and public clouds.
The OpenStack project was created to develop open source software designed to create and manage private and public clouds and to provide Infrastructure as a Service (IaaS). OpenStack is developing a set of interrelated software modules for managing large pools of computing resources (Nova), storage resources (Cinder, Glance, Swift) and network resources (Neutron), as well as modules for metering resources (Ceilometer), data processing (Sahara ) and "orchestration" (Heat). OpenStack implements the management of various virtualization technologies using a single portal and open APIs.
Now, more than 850 companies from 144 countries are participating in the OpenStack project, and new releases of OpenStack software are released every six months.
Using OpenStack “open” software as a basis for building a private cloud provides several important advantages. First of all, these are savings on licenses, compatibility with public clouds based on OpenStack and AWS, vendor independence and the broadest support in the IT industry, the ability to adapt the cloud to your needs thanks to the open source code, scalability due to the modular software architecture, and fast innovation by the whole OpenStack community. The openness of the OpenStack source code, a variety of functions and features for customization for individual needs entail difficulties in the design and implementation of the solution.
Building a private cloud based on it cannot be done as a simple plug-and-play procedure, for this, at least, you need to choose the correct configuration of the cloud hardware platform, otherwise it will not provide high performance of applications or it will turn out that some of the servers purchased for the cloud are idle . The implementation requires an appropriate expertise of the IT team, but most companies have no specialists who are well versed in the intricacies of the OpenStack technology, moreover, even during the implementation, there is no possibility to fully utilize the employee’s working time, and hiring is unreasonably expensive. As a result, the introduction of a private cloud based on OpenStack on its own is associated with high risks and can take several months, which is unacceptable for many companies that want to get the benefits of using cloud technologies as soon as possible.
Another problem is the support of a private cloud based on OpenStack, including software updates and troubleshooting. As in the case of implementation, for effective support requires a special expertise of IT professionals. To obtain expertise in this matter, the company will have to train its employees or hire qualified specialists to work. In both cases, this means significant additional costs that reduce the economic attractiveness of a private cloud. However, the availability of expertise in the IT team of the company does not mean that it can easily provide support for the cloud. The fact is that the OpenStack project develops according to the Open Source model, so the owner of a private cloud is responsible for managing new module releases and installing various patches, and if a problem is found in the work of some OpenStack module, the cloud owner can only ask for help OpenStack communities either try to fix the problem yourself.
As a result, if building a cloud is a supporting task for the company's business, a decision is made to involve a third-party organization to design, implement and support a solution on OpenStack. Moreover, any integrator will gladly tell you about his competence in this area and rich experience. But think about how many implementations a company could actually spend on a technology that was born only six years ago and is beginning to be seriously considered only recently? This question remains open and we, in turn, would like to talk about Fujitsu's experience in the field of OpenStack projects.
Fujitsu is actively involved in the work of the OpenStack, has the status of “Gold Member” and plans to enter the top three most active community members in the next two years. Moreover, it was decided to transfer all Fujitsu systems to its own K5 cloud, built on the OpenStack platform. Currently, a multi-year project is being carried out to transfer more than 640 of its business systems to K5, which run on 13,000 servers installed in Fujitsu data centers on different continents. The implementation of this project gives Fujitsu enormous experience in the design, implementation and support of OpenStack solutions, as well as in the migration of applications. A reduction in costs as a result of more than $ 300 million over 5 years. If desired, the customer can also take advantage of the benefits of K5 public clouds or deploy their virtual private cloud in one of the data centers of Fujitsu or partners.
Fujitsu's contribution to different releases of OpenStack
Fujitsu, together with Red Hat, developed the integrated PRIMEFLEX for Red Hat OpenStack system. This pre-tested for compatibility hardware and software complex minimizes risks when deploying infrastructure and helps avoid unpleasant surprises. A ready reference architecture, according to our estimates, reduces costs at the design stage by up to 40%. At the same time, not least, the support is fully provided by Fujitsu. In this case, Fujitsu is a single point of contact for support and software and hardware parts of the solution, which eliminates the element of shifting responsibility between different vendors. PRIMEFLEX for Red Hat OpenStack is built entirely on Fujitsu's own development and solutions from Red Hat and Brocade partners. PRIMEFLEX for OpenStack is based on PRIMERGY servers, which have repeatedly demonstrated the best results in performance testing for a virtual environment, and ETERNUS DX storage systems, representatives of a single line of storage systems that have full compatibility between systems of various classes and thus demonstrate unique scalability and preservation of investment.
PRIMEFLEX for Red Hat OpenStack Components
You can talk for a long time about the unique solutions of Fujitsu, but, as we have already discussed above, the main problem of OpenStack projects lies not in the technology plane, but in the plane of implementation and support services.
The Fujitsu service model is based on several service centers (GDCs) located throughout the world. And it is the merit of these centers that Fujitsu is among the five largest companies in the world in terms of services provided. However, few people know that one of these GDCs that provide services to the largest companies in the world is located in Russia in the city of Kazan. At the expense of the local resource, Fujitsu provides really high-quality consulting and development of the private cloud architecture, and can quickly implement and migrate business applications. Remote management and support services eliminate problems at the operational stage, and a well-constructed service desk operating in 24x7 mode will provide high-quality support to end users, allowing them to concentrate on solving their main tasks, rather than spending time communicating with administrators.
If you are seriously starting to look at the tasks of building a private cloud or migrating a part of systems to public clouds, along with integration with traditional systems in your data center, take a closer look at all the possibilities offered by the OpenStack platform. According to our estimates, this is the right choice Fujitsu has already made.
Cloud Approaches
There are two main options for switching to cloud computing:
- Building your own (private) cloud operating in a corporate data center;
- Using external provider services and translating applications to its public cloud.
Both approaches have their pros and cons. The advantage of a private cloud is high security (applications and their data do not leave the company), more effective control over the performance and reliability of applications, the absence of dependence on an external supplier and a virtually unlimited choice of virtual machine configurations. The public cloud requires significantly less initial investment in cloud technologies and allows you to pay for resources allocated to applications using the pay-as-you-go model, provides transparency in IT costs, and also significantly exceeds the private cloud in terms of capacity scaling and ease of management.
')
For successful implementation of a cloud infrastructure in a company, it must support both classic vertically scalable high-availability server applications that can work only on a single server, as well as horizontally-scalable applications designed for clouds that can run on multiple servers at the same time and do not require software high availability. The failure of one instance of this type of application does not lead to the failure of the entire service provided by the application.
Since both cloud variants have their pros and cons, the best application for a cloud technology company is a combined approach, in which some applications are deployed in a private cloud (usually business-critical applications and applications with increased security requirements), and all others in the public cloud of an external provider. Such a hybrid cloud usually combines both virtual and physical infrastructure, supports different service delivery models (IaaS, PaaS and SaaS) and different hypervisors (KVM, XenServer, ESXi and Hyper-V).
Hence, the task of ensuring the interaction between applications located in the private and public cloud, requires mechanisms that implement the exchange of data and application templates between these two clouds, and if necessary, transfer between them the applications themselves. Porting applications may be needed, for example, if you need to free up some of the resources of a private cloud due to a sharp increase in the load of one application. Therefore, some applications can be temporarily transferred to the public cloud. In addition, after testing in a public cloud a new version of the code of a business-critical application, it should be moved to a private cloud. Mechanisms are also necessary to allow, when changing the cloud service provider, to transfer all applications from the public cloud of the previous provider to its public cloud. Thus, we need ready-made solutions that guarantee the compatibility between private and public clouds.
The OpenStack project was created to develop open source software designed to create and manage private and public clouds and to provide Infrastructure as a Service (IaaS). OpenStack is developing a set of interrelated software modules for managing large pools of computing resources (Nova), storage resources (Cinder, Glance, Swift) and network resources (Neutron), as well as modules for metering resources (Ceilometer), data processing (Sahara ) and "orchestration" (Heat). OpenStack implements the management of various virtualization technologies using a single portal and open APIs.
Now, more than 850 companies from 144 countries are participating in the OpenStack project, and new releases of OpenStack software are released every six months.
What is the advantage of OpenStack?
Using OpenStack “open” software as a basis for building a private cloud provides several important advantages. First of all, these are savings on licenses, compatibility with public clouds based on OpenStack and AWS, vendor independence and the broadest support in the IT industry, the ability to adapt the cloud to your needs thanks to the open source code, scalability due to the modular software architecture, and fast innovation by the whole OpenStack community. The openness of the OpenStack source code, a variety of functions and features for customization for individual needs entail difficulties in the design and implementation of the solution.
Building a private cloud based on it cannot be done as a simple plug-and-play procedure, for this, at least, you need to choose the correct configuration of the cloud hardware platform, otherwise it will not provide high performance of applications or it will turn out that some of the servers purchased for the cloud are idle . The implementation requires an appropriate expertise of the IT team, but most companies have no specialists who are well versed in the intricacies of the OpenStack technology, moreover, even during the implementation, there is no possibility to fully utilize the employee’s working time, and hiring is unreasonably expensive. As a result, the introduction of a private cloud based on OpenStack on its own is associated with high risks and can take several months, which is unacceptable for many companies that want to get the benefits of using cloud technologies as soon as possible.
Another problem is the support of a private cloud based on OpenStack, including software updates and troubleshooting. As in the case of implementation, for effective support requires a special expertise of IT professionals. To obtain expertise in this matter, the company will have to train its employees or hire qualified specialists to work. In both cases, this means significant additional costs that reduce the economic attractiveness of a private cloud. However, the availability of expertise in the IT team of the company does not mean that it can easily provide support for the cloud. The fact is that the OpenStack project develops according to the Open Source model, so the owner of a private cloud is responsible for managing new module releases and installing various patches, and if a problem is found in the work of some OpenStack module, the cloud owner can only ask for help OpenStack communities either try to fix the problem yourself.
Fujitsu OpenStack Solution Expertise
As a result, if building a cloud is a supporting task for the company's business, a decision is made to involve a third-party organization to design, implement and support a solution on OpenStack. Moreover, any integrator will gladly tell you about his competence in this area and rich experience. But think about how many implementations a company could actually spend on a technology that was born only six years ago and is beginning to be seriously considered only recently? This question remains open and we, in turn, would like to talk about Fujitsu's experience in the field of OpenStack projects.
Fujitsu is actively involved in the work of the OpenStack, has the status of “Gold Member” and plans to enter the top three most active community members in the next two years. Moreover, it was decided to transfer all Fujitsu systems to its own K5 cloud, built on the OpenStack platform. Currently, a multi-year project is being carried out to transfer more than 640 of its business systems to K5, which run on 13,000 servers installed in Fujitsu data centers on different continents. The implementation of this project gives Fujitsu enormous experience in the design, implementation and support of OpenStack solutions, as well as in the migration of applications. A reduction in costs as a result of more than $ 300 million over 5 years. If desired, the customer can also take advantage of the benefits of K5 public clouds or deploy their virtual private cloud in one of the data centers of Fujitsu or partners.
Fujitsu's contribution to different releases of OpenStack
Fujitsu, together with Red Hat, developed the integrated PRIMEFLEX for Red Hat OpenStack system. This pre-tested for compatibility hardware and software complex minimizes risks when deploying infrastructure and helps avoid unpleasant surprises. A ready reference architecture, according to our estimates, reduces costs at the design stage by up to 40%. At the same time, not least, the support is fully provided by Fujitsu. In this case, Fujitsu is a single point of contact for support and software and hardware parts of the solution, which eliminates the element of shifting responsibility between different vendors. PRIMEFLEX for Red Hat OpenStack is built entirely on Fujitsu's own development and solutions from Red Hat and Brocade partners. PRIMEFLEX for OpenStack is based on PRIMERGY servers, which have repeatedly demonstrated the best results in performance testing for a virtual environment, and ETERNUS DX storage systems, representatives of a single line of storage systems that have full compatibility between systems of various classes and thus demonstrate unique scalability and preservation of investment.
PRIMEFLEX for Red Hat OpenStack Components
You can talk for a long time about the unique solutions of Fujitsu, but, as we have already discussed above, the main problem of OpenStack projects lies not in the technology plane, but in the plane of implementation and support services.
The Fujitsu service model is based on several service centers (GDCs) located throughout the world. And it is the merit of these centers that Fujitsu is among the five largest companies in the world in terms of services provided. However, few people know that one of these GDCs that provide services to the largest companies in the world is located in Russia in the city of Kazan. At the expense of the local resource, Fujitsu provides really high-quality consulting and development of the private cloud architecture, and can quickly implement and migrate business applications. Remote management and support services eliminate problems at the operational stage, and a well-constructed service desk operating in 24x7 mode will provide high-quality support to end users, allowing them to concentrate on solving their main tasks, rather than spending time communicating with administrators.
If you are seriously starting to look at the tasks of building a private cloud or migrating a part of systems to public clouds, along with integration with traditional systems in your data center, take a closer look at all the possibilities offered by the OpenStack platform. According to our estimates, this is the right choice Fujitsu has already made.
Source: https://habr.com/ru/post/319172/
All Articles