OpenStack

OpenStack is a solution that helps you create software-configurable data centers. It includes components for the organization:

• server virtualization,
• software-defined data storages,
• software-defined networks,
• "databases as a service,"
• cloud orchestrator.

As of autumn 2016, the OpenStack project is supported by more than 600 companies and about 63 thousand IT specialists in 186 countries of the world, including Red Hat, IBM, Intel, Huawei. Over 140 solutions have already been created on the basis of the platform. Among domestic companies, both private and government organizations are showing interest in open source software. For example, Sberbank considers an open platform as the basis for its IT systems. Experts attribute the growing interest in open source solutions to the high level of maturity of open source developments, their more flexible and rapid development compared to proprietary solutions, as well as the lower cost of implementing, using and developing such solutions.

In the summer of 2015, the OpenStack community includes 27,300 people, more than 500 organizations support the development of the platform. Over the years, over 20 million lines of code have been created by the efforts of the developer and user community.

Read about what cloud services can be obtained on OpenStack in the TAdviser special project Details.

OpenStack Developers

The OpenStack software product was developed by Rackspace for commercial provision of "infrastructure as a service" and for use in NASA scientific projects. Subsequently, the source codes were provided to the global independent community of developers, and the project continued to develop actively, and also became available for free private and commercial use.

More than 180 companies joined the development of the OpenStack project in November 2014. The highest level of membership ("platinum") in a non-profit organization among companies:

• AT&T,
• Canonical,
• HP,
• IBM,
• Nebula,
• Rackspace,
• Red Hat,
• SUSE.

In the world market, Red Hat, Mirantis, Enovance have the competencies for implementing solutions based on OpenStack, in Russia for 2014 ASD Technologies announced itself.

Advantages

In the world, OpenStack is actively used by large data centers, cloud and other service providers, telecom operators, media companies, government agencies and scientific organizations.

Open source solutions for creating private clouds can be attractive in several cases. First of all, they are of interest to those companies that have their own powerful data centers and want to implement flexible resource management in them according to the cloud model.

The second task force is companies that agree that the future lies with cloud computing, but for some reason they are not yet ready to entrust their data to the owners of public cloud services. By developing new systems for the private cloud, such companies will be able to transfer them to the public platform at relatively low costs in the future. Finally, open source private clouds are the best tool for experimenters to quickly and cost-effectively deploy a pilot platform and understand in practice how cloud computing is appropriate in the context of their own organization.

OpenStack is an open platform that helps create a cloud solution with a high level of reliability, avoiding the economic and political risks that users of proprietary products are exposed to.

OpenStack of Russia

In 2014, Russian business is very active in relation to this global trend, and interest in the OpenStack platform in our country is constantly growing. Several hosting providers offer OpenStack-based products and services, and a number of companies that use community member products to virtualize their own infrastructure. In particular, a number of OpenStack components are used by the company Yandex for several of its services.

2025

DSS Compatibility Driver Release

Aerodisk on March 28, 2025 announced the release of a driver that provides full compatibility of its DSS with the OpenStack platform, which greatly simplifies resource management and increases the flexibility of using DSS in various environments.

The driver works on all Aerodisk DSS implemented on the basis of the Aquarius platform. Read more here.

As part of RuBackup OneClick

RuBackup (part of Astra Group) on March 18, 2025 announced the release of RuBackup OneClick, a software system for quickly creating a data preservation system. The solution also included OpenStack. Read more here.

2022: Support with Softline's Coster solution

Softline April 22, 2022 announced plans to support cloudy infrastructures based on OpenStack within the framework of a multi-cloud platform. Coster Users of OpenStack solutions will have the opportunity to aggregate everything information about the costs of cloud infrastructure as soon as possible. More. here

2021: Expansion of integrations with Aerodisk DSS

On October 28, 2021, Aerodisk announced the completion of a set of works to expand the integration of its solutions with the OpenStack ecosystem of cloud infrastructure management tools. Read more here.

2018: OpenStack Queens release

On March 7, 2018, the next OpenStack release was released. 17th version of the cloud IaaS platform.

Many features have appeared in this release. These include support for virtual GPUs (vGPUs) and advanced container integration. In addition, several projects are included in the release, including the Cyborg hardware and software acceleration resource management system.

OpenStack Queens (2018)

vGPU

vGPU support appeared in the OpenStack Nova project and is a continuation of the work to create the Placement service that this project performed for several recent releases. This service is a dynamic inventory system in OpenStack that monitors the resources available for deployment. vGPU support extends to Nvidia Grid vGPU and Intel GVT (Graphics Virtualization Technology).

"The GPU can have thousands of cores, and if you try to only trivially transfer hardware resources to a virtual machine, the entire GPU will be there. Many companies want to have a fleet of different GPU servers that can be distributed between different tasks. Until now, there was no solution for this that would be easy to manage in a multiplayer environment. " Jonathan Bryce, Executive Director of the OpenStack Foundation

Cyborg

The Cyborg project came from the telecommunications world. It provides a universal framework for speeding up any processes, not just GPUs. Cyborg can use physical GPUs, user programmable gate arrays (FPGAs), and any other accelerators. OpenStack already had the Ironic Bare Metal service, which allows cloud operators to directly access hardware resources. Cyborg complements Ironic.

"Both are tools for direct control of equipment. Ironic is focused on server management in its entirety for acceleration, and Cyborg is focused on separate hardware functions. " Jonathan Bryce, Executive Director of the OpenStack Foundation

Containers

The Queens release also includes the popular open-source project OpenStack Helm for package management of the Kubernetes container orchestration system.

Helm, like many other projects, including Ansible, aims to provide OpenStack services with configuration and deployment management capabilities.

"OpenStack Helm and OpenStack Ansible are similar in that they are tools that should help manage OpenStack services, but they take different approaches." Jonathan Bryce, Executive Director of the OpenStack Foundation

Helm uses Kubernetes for lifecycle management and is designed specifically for containers. But some organizations are more comfortable using the Ansible configuration management tool.

LOCI

The LOCI (Lightweight Open Container Initiative) project is an alternative to the OpenStack Kola project, which represents a more complete approach to packaging a container image.

Path to Rocky

After the release of Queens, work begins on the next release, Rocky. It is likely to feature Fast Forward Upgrade, which will make it easier for cloud operators to update OpenStack releases. Fast Forward Upgrades will allow you to skip releases.

2017

OpenStack Pike Release

On September 1, 2017, the next OpenStack release was released. The 16th version of the cloud IaaS platform was called Pike.

It is focused on simplifying deployment and updating. Pike is based on the OpenStack Ocata released in March. The peculiarity of the release - the focus on stability and performance^[1].

Among the opportunities that have emerged:

• Nova Cells v2: Nova Cells computing architecture is required for large-scale deployments and interoperability between APIs. The second version of this OpenStack subsystem received database scaling tools, a simple text message protocol, the ability to disable domains with errors and eliminate individual points of failure.
• Python 3.5: Support for the latest version of the Python programming language is of paramount importance, since it contains the vast majority of OpenStack subsystems and APIs.
• etcd v3: a distributed system for storing etcd configuration parameters specified as a key/value pair is used as a low-level link for storing file content and coordinating consensus between nodes. It is being developed by the CoreOS team and has already found application, for example, Google in Kubernetes PaaS and in from. Huawei Starting with this release, etcd v3 will be used to store critical data.
• Ironic: This project penetrated the OpenStack computing environment along with the Juno release. One of the goals of Ironic is TripleO (OpenStack-on-OpenStack) or interchangeable complex applications without replacing the bare metal required to deploy the cloud. In the Pike release, the Ironic developers added the ability to connect to Neutron networks. Modular architecture with the ability to connect various devices allows vendors to organically integrate their solutions with OpenStack.
• Ironic, along with Cinder, Neutron, Nova and Swift, received support for the so-called rolling upgrades, allowing operators to add functionality without disabling cloud services, which will generally simplify the update of OpenStack.
• Cinder: The OpenStack block storage service has been enriched with the revert to snapshot function, which is required to restore the current state of the virtual machine from a previously saved one, as well as the ability to expand volumes. Returning to the saved snapshot can be useful after data corruption or resetting the settings after performing tests. The storage capacity expands without collapsing virtual machines, thereby maintaining the functionality of online applications.
• Kolla is a container lifecycle management tool that has also received some improvements. It provides ready-made containers and tools for deployment in the OpenStack cloud. The use of Kolla in conjunction with DevOps programs such as Kubernetes and Ansible simplifies the management and update of OpenStack.
• Swift: OpenStack object storage has learned to work with Erasure coding algorithms, which, by complicating the computational add-on, better protect data. Swift solves the problem of restoring information when the network falls using fragmentary data collection through separate data centers. The storage version has acquired a fast engine, allowing users to run several parallel processes on the server at the same time.
• Zun: the final release of OpenStack, an industry-ready container management service.

The main feature of Pike is the improvement of modularity. The modular architecture of OpenStack allows you to select specific functionality for connecting to the infrastructure stack - be it dedicated servers for creating clouds or backup storage of blocks without deploying a full stack of subsystems.

For example, OpenStack Ironic offers advanced integration for storing Cinder and Neutron blocks, and Cinder can act as stand-alone storage for virtual machines, dedicated servers or containers based on Docker or Kubernetes.

Thousands of large and small implementations in private and public clouds - this experience, gained over more than seven years, we transferred to Pike. Work was carried out on the errors, the composite OpenStack was finalized. It's time to talk about the maturation of the service for industrial use, but we also do not forget about its filling with new functionality. Jonathan Bryce, Executive Director, OpenStack Foundation

OpenStack Ocata release

On February 22, 2017, The OpenStack Foundation released the OpenStack Ocata release. There are no added features in it, it is focused on improving stability and performance.

The Ocata version came out as an exception earlier. This release is focused on improving the stability, scalability and performance of basic computing and network services^[2].

Presentation of Ocata Release Targets, (2016)

Ocata properties:

• Nova API for "hosting" computing, which helps users carefully allocate resources to the needs of applications. Cells v2 becomes the default choice for increasing Nova scalability;
• The OpenStack Horizon dashboard supports the Keystone federation of services as the user interface for the operating system profiler, which allows operators to identify performance problems with any OpenStack services;
• the federated Keystone identification service dynamically and automatically provides access to the project and, after successful authentication, assigns a role to the federation user;
• The reliability of the Ironic physical equipment service has been improved;
• Improved processor performance and utilization for Telemetry projects. Now, when using Ceilometer with the Gnocchi storage engine, you can save millions of indicators per second;
• Active/Active HA of the Cinder block storage service can now be implemented using drivers;
• The Congress management environment has improved the policy language. It now allows operations on network addresses, which improves network and security management.

Ocata has improved network-level support for its container-based application environment. With the help of the release, OpenStack services are available in containers. This simplifies the installation and management of OpenStack as a microservice application.

Added container platform application environments and deployment tools. These include Kolla (OpenStack services enclosed in containers), Kuryr (combines the transfer of services over networks with their storage) and Zun (container management).

A year ago, people thought containers meant the death of OpenStack. But in reality, they have become a source of momentum for the cloud platform. Using container orchestration environments to run OpenStack as an application makes it easier to manage. This means fewer people are needed to maintain the cloud, making OpenStack a more affordable and practical option for small cloud owners. Jonathan Bryce, Executive Director of the OpenStack Foundation

The nova-status upgrade check command has been added, it allows operators to check the readiness of their systems and see if they can be safely updated to Ocata. If not, you will be told which issues need to be resolved before upgrading.

2016: HPE and Mirantis cut the number of OpenStack developers

In the fall, it became known about major reductions in specialists in two leading companies in the development of the free cloud platform OpenStack: Hewlett Packard Enterprise (HPE) and Mirantis.

Last year, 2015, HPE and Mirantis employees occupied the first two lines in terms of the number of changes made to the OpenStack codebase. According to new statistics, the contribution from them this year has decreased significantly. This trend is also consistent with recent information about personnel changes in both companies. In particular, Canonical founder Mark Shuttleworth said in an interview with ComputerWorld UK last week that HPE had fired "the entire OpenStack team." An HPE spokesman confirmed that the company is undergoing restructuring among OpenStack specialists, but did not name the exact numbers, noting that Shuttleworth's statement was exaggerated.

According to Boris Renski, one of the founders of Mirantis, "less than 100" engineers were cut in the company. This process began after the acquisition of TCP Cloud, which specializes in selling private cloud infrastructure services based on OpenStack. This takeover led Mirantis to the need to "balance between a culture of exceptional software engineering and expertise in operation." The cuts affected two groups of engineers: the developers of the Fuel project (a simple OpenStack installer), which is planned to be revised towards the IaC (Infrastructure as Code) approach, and the authors of the internal PaaS project, the work on which is shifting in favor of supporting the Kubernetes container orchestration platform. In his letter, Renski emphasized that the company "does not leave OpenStack and does not leave the OpenStack distribution business" ^[3]

2015

OpenStack 11 Kilo release released

On May 2, 2015, the OpenStack community unveiled the eleventh release of the OpenStack Kilo platform. About 400 improvements have been made in this release^[4]

1492 developers from 169 companies took part in the development of the product. When preparing changes for release, the companies most actively showed themselves:

• Red Hat,
• HP,
• IBM,
• Mirantis,
• Rackspace,
• Yahoo!,
• NEC,
• Huawei
• SUSE.

On May 2, 2015, about 500 companies and about 24,000 independent developers joined the development of OpenStack. Cloud deployment products on the OpenStack platform are being developed by companies:

• Rackspace,
• Red Hat,
• SUSE,
• Canonical,
• IBM,
• Cisco,
• Dell
• HP.

As of May 2, 2015, the OpenStack platform includes a set of interconnected components focused on creating public cloud services, such as Amazon EC2 and private clouds, running on their own equipment and fully controlled by the company. Supported virtualization systems include:

• KVM,
• QEMU,
• Xen,
• Hyper-V,
• Citrix XenServer,
• LXC containers
• containers VMWarevSphere ESX//.ESXi

OpenStack scales well and is capable of servicing an infrastructure of hundreds of thousands of virtual servers.

OpenStack Interoperability Diagram, 2015

As part of the updates:

• OpenStack Object Storage (Swift) - distributed, fault-tolerant and highly trusted object storage. Stored objects are stored simultaneously on multiple cluster nodes in the datacenter, providing automatic integrity control and replication when nodes are disconnected/added. The storage is scaled horizontally, that is, to increase the size, you just need to add new nodes, which are configured automatically. When a node fails, its contents are reproduced on other network nodes to ensure proper redundancy. Duplication of information allows you to use typical low-cost servers to form a cluster, without caring about the reliability of each of them separately.
  • Support for redundancy code (Erasure Code, EC), designed to restore the contents of data blocks based on redundant information stored in other blocks. Compared to duplicate replication, redundancy codes make it possible to dispense with smaller storage, but require more computing and network resources to work. Of the applications for which this method is well suited, storage in one region of large data sizes is mentioned, which is accessed quite rarely.
  • Implementation of composite tokens, which allow OpenStack services on behalf of the client to store their data in Swift, so that individually not the client and not the service, cannot update this data without the consent of both parties. For example, a user can initiate a virtual machine snapshot through Nova, Nova will send this request to Glance, which will record the image in the Swift container in the form of a set of objects. Then the user will not be able to update the image without receiving a token from the service, and the service will not be able to update the image without receiving a token from the user.
  • The ability to take into account the weight of the device when placing data, which allows you to organize the addition of new zones or regions to the cluster without initiating the instant movement of large amounts of data, and also leads to more efficient processing of balanced clusters in which one zone contains significantly more data.
  • Improve data replication between regions. Only one replica moves in one pass, which gives the remote region the ability to perform internal replication without moving unnecessary data through an external network link.

• OpenStack Compute (Nova) is a toolkit that allows you to automatically create and manage large groups of VPS servers. The system is based on NASA Nebula technology and the developments used to provide the Rackspace Cloud Servers service. OpenStack Compute is written in Python using the Tornado and Twisted frameworks, the AMQP protocol is used to exchange messages, and the Redis database is used to store data.
  • A new Nova API (2.1) has been implemented, which is backward compatible with API 2.0 and is notable for the introduction of a system for adding innovations through the microversion mechanism. API 2.1 also provides support for x509 certificates.
  • Support for live updates if the database storage scheme needs to be changed.
  • Support for second-generation Hyper-V virtual machines and SMB partitions has been added to the Hyper-V driver;
  • The Libvirt (KVM) driver has the ability to attach a vCPU to a guest system, support for SMB-resdels and support for iSCSI initiator QEMU.

• OpenStack Image Service (Glance) is a registry of virtual machine images that allows you to register new virtual machine images and ensure their delivery to the desired nodes before execution. Provides a storage of virtual machine images presented in a wide range of formats: Raw, AMI, VHD (Hyper-V), VDI (VirtualBox), qcow2 (Qemu/KVM), VMDK (VMWare) and OVF (VMWare). To unify operations for obtaining information, adding and delivering virtual machine images, a special Glance REST-ful API is used;
  • The ability to deactivate the image;
  • Support for image conversion during the import process;
  • Choice of hashing method in addition to SHA-1;
  • Support for Metadata tags;
  • Experimental catalog indexing API;
  • Support for rebooting configuration files when receiving a SIGHUP signal;

• OpenStack Identity (Keystone) is a package for unifying authentication tools and ensuring the integration of OpenStack components with existing authentication systems, including those built on the Active Directory and LDAP platforms . It is also possible to use ordinary login tools using a login/password or AWS. Keystone has operations such as managing users, projects and access rights;
  • Federated identification is enhanced to work in hybrid systems with multiple cloud environments.
  • Implementation of hierarchical linking of project settings;
  • Support for Fernet tokens;

• OpenStack Dashboard (Horizon) is a web-based system management interface based on the Django web framework and the django-openstack application. The interface is presented for both administrators and users. A wide range of tools is supported for managing resources, creating and launching environments, setting limits. Plugins can be connected, for example, with the implementation of monitoring tools.
  • Implementation of themes for changing the appearance;
  • Single-Sign-On authentication support;
  • The interface for managing the data processing system (Sahara) has been radically redesigned;
  • Launch Instance Wizard added

• Networking (Neutron) is a framework for performing tasks related to creating, configuring and maintaining networks within data centers. Quantum supports dynamic network configuration and can be used to configure both virtual networks and physical hosts. Quantum supports the extension of functionality through plugins and can solve various administrative tasks, from creating ports to configuring routes and VLANs.
  • Added VLAN support to DVR in addition to VXLAN/GRE;

• Brocade Vyatta vRouter Firewall VPN Intel * New Drivers: A10 Networks LBaaS V2, Brocade LBaaS V2, Brocade ML2 for MLX and ICX Switches,, Brocade Vyatta vRouter, Cisco CSR VPNaaS, Freescale FWaaS, Mcafee NGFW VPNASPAGASA
  • New plugins: L3 routing on Brocade MLX, Brocade Vyatta vRouter L3, Distributed Virtual Router L3 switches based on Dragonflow SDN.

• Block Storage (Cinder) is a multifunctional and expandable storage of block devices and disk partitions, which also supports work with previously created partitions deployed during the use of past releases of OpenStack. Previously, Cinder was implemented as a subproject to the OpenStack Compute service, but from now on it is allocated to a separate subsystem developed by a separate development team.
  • The ability to change the database schema without restarting;
  • Support for adding/removing partitions from an existing consistent group;
  • Possibility of creation of an agreed group from the snapshot of another agreed group;
  • The backup service can now be applied to encrypted partitions;
  • The ability to create private partition types;
  • To ensure fault tolerance, you can now bind a partition to multiple instances of OpenStack Compute;

• OpenStack Metering (Ceilometer) - aimed at collecting data on the operation of the system and monitoring, providing users with a single source of data on the use of all OpenStack services;
  • Support for Ceilometer API RBAC;
  • Support for receiving metrics through Gnocchi

• OpenStack Orchestra (Heat) - provides a service for managing the lifecycle of cloud-based applications and automating distribution for applications of computing resources, network bandwidth and storage space.

• OpenStack Database Service (Trove) is a system for maintaining a cloud database service (Database as a Service), based on both relational and non- relational DBMSs. Trove simplifies the maintenance of multiple DBMS instances by providing tools for operations such as DBMS deployment, configuration, patching, backup, disaster recovery, and monitoring. Full support is provided for MySQL and Percona. Experimental support is available for MongoDB, Redis, Cassandra and Couchbase;
  • Support for a new replication strategy based on MySQL 5.6 asynchronous GTID replication;
  • Vertica, Vertica Cluster, DB2 and CouchDB storage support;
  • Control API extension;

• OpenStack Data Processing (Sahara), a component for automating the maintenance and management of large data processing clusters based on Apache Hadoop or Apache Spark.
  • New plugins for MAPR, Apache Storm and Apache Hadoop 2.6.0;
  • Added event log maintenance

• Iron (Bare Metal Provisioning Service), a new component that allows you to use existing methods of deploying virtual machine configurations to deploy systems on real hardware or for applications such as Linux containers.

Cisco and IBM bought OpenStack developers

In early June 2015, Cisco and IBM announced acquisitions through which American corporations can simplify the implementation of the private cloud platform OpenStack in enterprise IT infrastructure, writes PCWorld.

Cisco bought the startup Piston Cloud Computing, which develops CloudOS software designed to build open source cloud infrastructures based on OpenStack to provide IaaS services. Features of this product are automatic installation, high security and compatibility with public clouds based on OpenStack.

Cisco, IBM to simplify IT adoption of OpenStack

It is noted that Piston solutions will strengthen Cisco's cloud business and add new practical experience with the underlying infrastructure underlying Cisco OpenStack Private Cloud.

The takeover of IBM was the provider of cloud services Blue Box. It provides the OpenStack platform in the form of an Internet service that does not require installing and configuring software on client computers. At the same time, the platform itself is flexible and scalable, noted in Blue Box.

The cost of acquisitions made by Cisco and IBM is not specified. Also, corporations did not talk about ways to integrate the technologies of purchased firms with their products.

Experts are confident that through the announced deals, IBM and Cisco will be able to ensure a large-scale implementation of the OpenStack platform in the corporate environment. However, the growth of this market is in question. In April 2015, Nebula, a company specializing in OpenStack, closed.

greatly disappointed that the market is likely to take several years to mature," Nebula said, commenting on the termination of its operations.

Several dozen Nebula specialists went to work with Oracle. Thus, the American corporation has shown its high interest in the OpenStack project. HP and EMC are also interested in him. The latter in 2014 bought Cloudscaling (developer of IaaS OpenStack). It is worth noting that fewer and fewer small manufacturers of OpenStack solutions remain on the market.^[5]

2014: OpenStack Icehouse release

On April 17, 2014, the next update of the OpenStack platform, Icehouse, was released.

It has 350 not entirely fundamental innovations, made taking into account the wishes of users, mainly, and 2902 adjustments. In fact, Icehouse is a kind of embodiment of feedback from developers with the global community. The largest amount of work on writing Icehouse code was done by engineers from Red Hat, IBM, HP, Rackspace, Mirantis, SUSE, OpenStack Foundation, eNovance, VMware, Intel, Samsung, Yahoo and Comcast. Assistance in the development of Icehouse was provided by 1202 participants, which is 32% more compared to support for the previous version of Havana, published six months ago.

The main attention of the developers focused on stability, testing and development.

List of major changes:

• The OpenStack Compute (Nova) module simplifies the installation of updates and enhances the requirements for third-party drivers.
• OpenStack Object Storage (Swift) significantly increases the visibility of many functional components, which in addition are available through program calls.
• OpenStack Block Storage (Cinder) provides incremental server migration in heterogeneous environments.
• OpenStack Networking (Neutron) is tightly consolidated with other OpenStack modules, which increases the reliability of the network and the consistency of its elements.
• OpenStack Identity Service (Keystone) supports federated authentication, OpenStack Orchestra (Heat) automatically scales additional resources across all platform subsystems, and the new configuration API simplifies application lifecycle support.
• OpenStack Telemetry (Ceilometer) offers access to automatically collected data (for example, billing information).
• The OpenStack Dashboard (Horizon) has been expanded with new navigation elements and support for 16 national languages, including Serbian and Hindi.

According to analysts at Forrester Research, OpenStack will become de facto another IaaS standard by the end of 2014.

2013

OpenStack Grizzly release

The new release of the OpenStack platform has the ability to manage much larger clouds than before. Grizzly improves network and storage management over the previous version of Folsom.

With Grizzly, you can manage much larger computing resources: tens of thousands of server nodes. The release introduces Nova Compute Cells, which allows you to scale very large systems distributed across multiple locations without using databases and clustering message queues. The release includes a NoDB host architecture that reduces dependence on the central database.

Changes to virtualization management support more features that use multiple hypervisors (ESX, KVM, Xen, and Hyper-V).

Grizzly also added protocols for distributed storage and network features such as hot add and replace network devices, in which all changes are made without any delays or interruptions.

Administrators can now set quotas to manage the growth of object-oriented memory environments, and perform large batch operations on that memory, making it easier to deploy and manage large clusters. The new release also adds the ability to manage heterogeneous environments for storing data through a single access point. At the same time, end users can now allocate storage in such a way that the required level of performance, efficiency and costs are met.

Improvements to the network subsystem include support for technologies such as Big, Switch, Hyper-V, PlumGrid, Brocade and Midonet, complementing the existing support for Open vSwitch, Cisco UCS/Nexus, Linux Bridge, Nicira, Ryu OpenFlow and NEC OpenFlow. Added more scalability and increased availability by distributing L3/L4 and using dynamic host configuration protocol services across multiple servers.

Grizzly OpenSrack Identity has added a new token format based on standard PKI functionality, allowing offline authentication through tokens without service calls to Identity.

All new computing, storage and networking features are available through the updated OpenStack Dashboard, which has added support for more languages.

OpenStack Havana release

As a result of voting in early 2013, a name was chosen for the new version of the OpenStack cloud operating system. It will be named Havana after the intercommunal territory in Oregon.

The release will follow Grizzly (before that there were Folsom, Essex, Diablo, Cactus, Bexar and Austin). The codenames of the OpenStack versions traditionally reflect hints of settlements in which the company's summits are held. For example, Grizzly is a hint of a bear on a California state flag. The next summit is scheduled for April 15-18, and will be held in Portland, Oregon.

The current version of the scalable cloud operating system OpenStack is Folsom, released in January 2012. It included support for network management, block storage scheme and Hyper-V. In September of that year, an update was released to Folsom. Grizzly is still under development, its official release is expected on April 4, 2013 (a release candidate will be available in mid-March).

The open IaaS project OpenStack is supported by the largest companies aimed at the corporate market. These include AMD, Intel, SUSE Linux, Red Hat, Cisco, Dell, HP, IBM, NEC and VMware. The goal of the project is to provide an easily implementable and scalable solution for public and private clouds. This is achieved using a wide variety of APIs.

Dell supports Windows Server Hyper-V in its OpenStack platform

Dell announced on May 27, 2013 that it supported Microsoft WindowsServer Hyper-V as one of the possible hypervisors in its OpenStack cloud platform. For the first time, the Windows Server Hyper-V hypervisor on the OpenStack platform is offered for private clouds. This will give customers additional flexibility and choice when performing an OpenStack load on an existing Windows Server environment.

For OpenStack, Dell will support key features of the Windows Server 2012 virtualization platform, including network virtualization, policy-based isolation, Quality of Service (QoS), and multi-tenancy. In addition, Dell is working with other OpenStack developers to optimize OpenStack components, including OpenStack Compute for Hyper-V, OpenStack Block Storage for Windows, and Dell Crowbar. It is assumed that the basic functionality will be implemented in the future version of the OpenStack project.

"For almost three years Dell now, he has been an active sponsor of OpenStack," said the Forrest Norrod CEO of Dell Server Solutions. "This project Dell will help the OpenStack community gain additional investment benefits Windows Server 2012 with the ability to further explore OpenStack-based solutions and workloads in existing environments."

2012

Creating The OpenStack Foundation

In September 2012, after almost a year of preparations, he began working as an independent non-profit organization, The OpenStack Foundation. Now OpenStack, an open source software stack for creating a cloud infrastructure as a service, is free from the control of hosting provider Rackspace. The company limited the development of the OpenStack developer community for the first two years of the project, but last year its leaders decided that open source software should develop independently of RackSpace. The newly created alliance was taken over by community leadership and the OpenStack brand.

Participation in the alliance is distributed so that no one has too much influence in the governing council, according to OpenStack. Among the "platinum" sponsors of the project are AT&T, Canonical, HP, IBM, Nebula, Rackspace, Red Hat and SUSE, among the "gold" ones are Cisco, Dell, NetApp and Yahoo, and soon Intel, NEC and VMware will join them.

Microsoft Hyper-V Opt-Out Discussion

The OpenStack cloud platform community is considering not supporting the Microsoft Hyper-V hypervisor . Tieri Carré, responsible for managing OpenStack releases, said that Hyper-V code is planned to be excluded from the currently being developed version of the OpenStack Compute computing resource controller, since this code is "deliberately defective and not supported by Microsoft itself ."

Seizing the code will presumably not create any problems, since no one uses Microsoft's hypervisor in OpenStack-based projects, says Joshua Mackenty, a former technical architect for NASA's Nebula platform, which became the basis of OpenStack. According to him, Microsoft hired a contractor to integrate Hyper-V with OpenStack, but "he never finished this work."

Microsoft responded to the news of the possibility of excluding Hyper-V support code from OpenStack by stating that "the corporation intends to work with the OpenStack development community to fix existing problems with this platform and Hyper-V."

OpenStack Essex Release

April 2012: Essex includes over 150 new features. The information published on this project states that more than 200 developers participated in the development of the new version. Among the main improvements are the following:

• OpenStack Compute (codenamed Nova) - improved code stability and integration with Dashboard and Identity components, provides equal support for type 1 hypervisors, which will help create conditions for users to work transparently with them. In addition, support for authorization and migration functions has been improved without stopping services in a network with many hosts. Support for high-performance computer computing was also provided, new options were added for working with block storage, including for storage systems from Nexenta, SolidFire and NetApp;

• OpenStack Object Storage (codenamed Swift) - new features have appeared that simplify compliance and data protection, as well as the ability to specify the lifetime of objects in accordance with document retention policies, increased resilience to data failures and corruption, and improved disaster recovery functionality. In addition, there are new features that are valuable for service providers, including the functions of downloading data directly from an authorized web page, limiting the maximum number of containers for one account;

• OpenStack Dashboard (codenamed Horizon) - the first full release of OpenStack Dashboard provides administrators and users with the ability to access and manage cloud resources using a separate portal. It allows you to easily connect and use third-party products and services, such as monitoring services;

• OpenStack Identity (codenamed Keystone) - the first full release of OpenStack Identity provides a single authorization system for all major elements of the cloud operating system. It provides the ability to authorize, including based on login/password, tokens and AWS;

• OpenStack Image Service (codenamed Glance) - a virtual machine image management service has received several key updates that improve its usability, improve authorization mechanisms and image protection.

More than 155 companies, including NASA, Rackspace, Intel, Dell, Canonical-Ubuntu, HP, SuSE, support open cloud architecture.

OpenStack Folsom Release

On September 27, 2012, the Folsom release of the OpenStack cloud platform was released with new networking and storage features, as well as support for Hyper-V, and the next version of Grizzly is expected in March 2013.

Folsom includes more than 185 new features and improvements, which have been worked on by over 350 open source developers.

'Since the release of Essex in April 2012, the developers have created a lot of new things. The most important thing, of course, is that Quantum is now a basic project and virtual network services are already on the way, "writes Emilien Macchi, a French OpenStack enthusiast who practiced at StackOps. - And for the volume service, there is now its own separate Cinder 'project.

Here's what Kevin Jackson, one of the creators and architect of the OpenStack infrastructure, reported about Folsom: 'Speaking of OpenStack Folsom, one cannot fail to mention two notable new projects: Quantum and Cinder. Quantum provides a functionally rich and extensible API for software-configurable networks. It allows you to specify much more diverse network topologies than could be done so far, for example, configure multi-level applications and allocate subnets. Quantum also allows you to build more complex configurations of the main node using plugins, for example, using excellent Open vSwitch software or creating QoS and security functions. Quantum enables OpenStack users to fully control all aspects of their cloud computing environment without compromising the underlying infrastructure and security of the underlying OpenStack environment. In fact, Quantum creates the conditions for a real collective lease without any restrictions on virtual local networks. Cinder is an OpenStack block storage project formerly known as nova-volume. Separating this important group of basic functions into an independent project will allow them to be developed independently of Nova's computing functions. For users, this means faster development and error correction and new features that meet modern hosting requirements. Among the important improvements to Folsom and the resumption of support for the Microsoft Hyper-V hypervisor.

This is the sixth OpenStack release in more than two years of its history. The newly created OpenStack Foundation expects to release its next release in six months, codenamed Grizzly.

The main developers of the Grizzly project work in Rackspace, SwiftStack, Nebula, Nicira (owned by VMware), SolidFire and Red Hat.

Gartner criticizes OpenStack project

Analytical Gartner published in the fall of 2012 critical analysis of a cloud platform with the open OpenStack code. Developers of the platform - the Rackspace company and other known providers of cloud services - are guided by own commercial interests, despite openness of the source code. Many companies which expressed support of OpenStack in fact do not allocate essential resources for its development, and the fragmentation arising because of collision of interests of various participants of the project reduces compatibility of cloudy systems on the basis of OpenStack, the analyst of Gartner considers.

Supporters of OpenStack in reply point to recent creation of the independent organization OpenStack Foundation which will coordinate development, release of the next version of the platform (Folsom) and fast opening of the OpenStack Summit conference where the directions of its further development will be discussed.

OpenStack is not the standard standard, the analyst and as the number of the companies specializing in service of systems on the basis of OpenStack while it is small, the risk of a binding to the supplier when using OpenStack is not lower, than when using commercial cloud platforms emphasizes. OpenStack, od­na­ko, uka­zy­va­yut a byst­ryy increase in the number of project uchast­ni­kov.

OpenStack, an open platform for managing cloud systems, has been widely touted as the future of the entire cloud infrastructure, both public and private, but in fact it is an evolving project, the future of which, although promising, is still uncertain.

Advertising implications:

• The hype around open cloud management platforms (CMP) leads some users to draw unreasonable conclusions that can lead to incorrect decisions when choosing a cloud management platform to create a private cloud, or when buying a cloud infrastructure as a service (IaaS) from a service provider.
• The marketing activities of supplier companies make Chief information officers believe that OpenStack is a stable, developed platform ready for widespread distribution. In fact, this project is at an early stage of its development, has unstable source code and a limited set of functions.
• Chief information officers may confuse adoption API with adoption of a Cloud Management Platform (CMP) that supports this API. Which forces them to limit the choice of technology platforms to their detriment.
• Supporters of OpenStack argue that it will eventually be able to supplant full-featured CMP platforms in commercial enterprises, in particular VMware vCloud Suite. But currently, CMP OpenStack offers only basic CMP capabilities.
• OpenStack can eventually become a reliable open core of a large and successful ecosystem of commercial offerings.

Recommendations:

• Do not assume that "open source" is the equivalent of open standards, broad interoperability, and freedom from commercial interests. In fact, OpenStack is dominated by the interests of vendor companies that want users to use their own offerings, and perhaps include a vendor binding.
• Evaluate OpenStack impartially as any other vendor solution, and remember that in its current state, it is most suitable for beginners with significant technical resources, high risk tolerance, and the need for a low-cost, high-end cloud infrastructure.
• If interoperability is a problem, use a third-party cloud API library or cloud management service that supports multiple cloud APIs, CMPs, or service providers.
• Do not make plans for the future development of your data center (data center), which should be based on OpenStack. Although this is a promising project, there is no guarantee in its future success.

OpenStack Analysis:

OpenStack is an open source CMP platform that provides some of the basic capabilities that are needed for the IaaS cloud service, including virtual infrastructure management and offline service delivery. It is not a single structure and is divided into several components, while users can select only some components at the same time.

Implications and Top OpenStack Ad Recommendations and Reality

The original and most important components are data processing services, Nova (similar to Amazon's EC2 service) and object data storage, Swift (similar to Amazon's S3 service). In addition, over time, other loosely connected components were added, including Quantum (network controller), Keystone (identity and access control), Glance (virtual machine image directory) and Horizon (self-service portal).

OpenStack source code is provided under the Apache 2.0 license, and there are also commercial distributions. Intermediate releases are released every six months; each release is alphabetically assigned a letter code, starting with "Austin" in October 2010 and ending with "Folsom," expected in September 2012 (see Table 1).

Table 1. Key OpenStack Projects

The Swift project appeared as software for the Rackspace Cloud Files service. This is the most stable and widely distributed part of the OpenStack project, with a successful history of large-scale use, which is usually used by IaaS cloud service providers, and has become the simplest implementation of object data storage such as the S3 service, but not compatible with S3.

Nova was created at NASA, as a result of overwriting the code of the Eucalyptus platform written in the Python programming language as open source CMP in Java EE, compatible with the Amazon Web Services (AWS) EC2 API. While NASA was developing Nova, Rackspace, a provider of administered hosting and cloud services, was working on a successor to the technology used in its proposed Cloud Servers service, which was obtained through the purchase of Slicehost in 2008. NASA and Rackspace decided to join forces in an open source software project (see Note 1). The consolidation of previous developments, which required a significant amount of work, created a new and unstable basis for the source code. The Nova project has a short history of mass production and operation. Currently, the first users of OpenStack are organizations with a strong technical base and a high risk tolerance. As a rule, these are service providers, financial companies and research institutions. Most of them either seek help from consultants or invest significant technical resources in the project.

Two key service providers are in the process of widespread adoption of Nova technology. Rackspace's Nova-based Cloud Servers service became public in August 2012. New customers will be offered Nova-based servers, and existing customers will be able to run new Nova-based virtual machines (VMs), but in essence this will lead to the gradual expansion of their Nova-based cloud. Rackspace will serve two platforms simultaneously until it can seamlessly move existing customers to the new platform, but that means it won't achieve instant mass use of Nova. HP offers Swift-based solutions, but its Nova-based solutions are in beta mode, and will only become publicly available at the end of 2012, so it will not achieve mass use either.

the past, Rackspace played a major role in the OpenStack project, but it will move on to an organization where voting rights will be determined mainly by the company's financial contribution. Participating companies hope that combining their efforts will quickly expand the technical capabilities of OpenStack. Their goal is to provide an alternative to AWS and VMware vCloud services. They believe Amazon and VMware are much more of a threat to them than they are.

Gartner noted the growing number of information requests for the OpenStack platform from Chief information officers, technicians, suppliers, service providers and investors. We conclude that active advertising of the OpenStack platform has created a number of dangerous myths that are reflected in vendor, partner, and investment decisions, including the arbitrary choice of the OpenStack platform without considering alternatives or performing a technical assessment. This study looks at the impact of advertising and also provides recommendations based on the practical realities of OpenStack as of today.

Impressions and recommendations:

The hype around open source CMP platforms leads some customers to draw unreasonable conclusions, which can lead to incorrect solutions to choose the suppliers of the CMP platform for creating a private cloud, or the purchase of the IaaS cloud service from a service provider.

Some people have been led to believe that because OpenStack is open source, it is an open and widespread standard, with broad functional compatibility and freedom from commercial interests. In fact, OpenStack is guided by commercial interests, since for related companies it is a business development strategy, and not an initiative of the community of altruistic individual sponsors. Some participants, especially Rackspace and other service providers, are afraid of the growing influence of AWS in the IaaS cloud services market and are not sure that they can, on their own, collect the engineering resources necessary for successful large-scale competition with AWS, and do not want to pay a permanent license fee for commercial CMPs such as the VMware vCloud software kit.

Rackspace and HP believe that with OpenStack they will be able to offer hybrid public/private solutions if they can achieve the use of OpenStack in internal enterprise clouds. Other companies, such as Piston Cloud Computing, Nebula and Cloudscaling, provide commercial distribution of OpenStack, along with professional service and support. Many companies, such as Cisco, Citrix (see Note 2), Dell and Red Hat, want to ensure that OpenStack is compatible with their products, as well as limit the growth of VMware's market influence.

OpenStack has attracted a large number of companies offering source code development, but many companies offer only specialized code, such as drivers for their hardware and prevent the creation of interoperability that requires common platforms that could facilitate switching between suppliers. The interests of companies also have a strong influence on the choice of features to add and how to implement them, as each company wants each feature to work best with its own products.

The OpenStack platform is not an open standard because it provides neither interoperability nor mobility. However, it has an open community and everyone can take part in discussions and offer additions to the source code. The technical manual is chosen by the participants, although most of the participants are employees of financing vendor companies, and in practice, the technical manual is carried out by employees of vendors. Voting rights on the board of directors of the future organization will be determined mainly by the sponsorship of vendors. The open nature of the community, combined with closed meetings of sponsors, sometimes leads to strong internal confrontation between community members, especially during the work of Rackspace to create a non-profit organization OpenStack Foundation.

There are problems with interoperability between versions of OpenStack, due to the fact that the project is in the stage of emergence and constant change, although these problems decrease as the project develops and compatibility becomes a community priority. There are also problems with interoperability between distributions, especially when you take into account that commercial distributions of the OpenStack platform generally include private extensions that reduce the level of integration.

In addition, since OpenStack is essentially a component platform rather than a product, some distributions may use components other than the basic version of OpenStack. For example, instead of a Swift project, another open source Ceph project may be used for storing data, or they may not include certain components and related functionality at all. These differences between versions and distributions extend to APIs, making it difficult for third-party libraries and tools to reliably support OpenStack. This lack of interoperability, even within OpenStack, will also limit the ability to build fully open hybrid public-private or federated cloud systems; instead, they will most likely require access to the services of specific service providers and the use of distributions they support.

As vendor companies seek to commercialize the OpenStack ecosystem, there is a notable trend of being tied to a vendor. For users, there is no big difference to whom to bind, to OpenStack or licensed CMP. In fact, since OpenStack is not yet a generally accepted standard, the number of solution providers is very limited (whether hardware, software or services), which further narrows user choice than the adoption of VMware vCloud. There are also few implementations of OpenStack CMP. In practice, AWS APIs have the widest ecosystem, including a large number of CMPs, both open and private, offering compatible implementations.

Recommendations:

• Evaluate OpenStack with no less care than any other licensed commercial offer.
• Do not consider OpenStack as a preferred standard when determining API and CMP compatibility in the context of selecting technology or service providers unless you have a business or technical reason to do so.
• Government agencies should be especially attentive and not accept the OpenStack platform for a generally accepted open standard. This is especially important during the development of vendor selection requirements. Since there is only one implementation of the OpenStack CMP platform, there are few OpenStack service providers and even fewer alternative ecosystems (in terms of hardware, software management, tools and libraries), due to the lack of a significant market share, OpenStack currently limits rather than increases choice.

The marketing activities of supplier companies make Chief information officers believe that OpenStack is a stable, developed platform ready for widespread distribution, while in fact this project is at an early stage of its development and has unstable source code and a limited set of functions.

Swift's release has successful experience and well-established code, but the rest of the OpenStack releases are completely new. OpenStack has had problems with source code stability in the past, although recent changes in the development process should increase the level of stability of future releases. However, it still has difficulties with project management and scope management, and thus jeopardizes the timely development of planned functions and the overall development of the project.

Quality control (QA) has improved over time, but remains a challenge. Vendor companies and service providers have developed the best quality control systems and testing tools, but do not submit them to the open source project, as they can be a competitive advantage in creating service offers and reduce the time to market distributions and related software (software).

As a rule, vendor companies involved in the development of OpenStack believe that it will be able to achieve a level of stability and technological readiness comparable to standard commercial enterprise-class software no earlier than the end of 2013. In addition, after each new release, it takes several more months to achieve stability of the source code, since the fixes are applied to all new code that was released with the release.

The projected timeframe for achieving stability is constantly changing. Usually, the developers believe that this takes another one and a half to two years, but they say this for several releases, and now there is no specific plan to determine the timing of stability. Typically, commercial software reaches stability by this time of its development, and the estimated time period is too long, even for complex open source projects.

Ongoing changes to the project also make it difficult for individual organizations - whether customer organizations or vendor companies in the OpenStack ecosystem - to create and maintain extensions to open source. For example, one of the advantages of using open source is that you can add any missing functions yourself, but if the way the function is implemented is incompatible with the next version, you will have to change and, possibly, completely rewrite your code. Due to the fact that OpenStack currently has a very limited set of functions, many organizations are forced to supplement OpenStack so that it can meet their requirements, so this is a serious problem that can reduce the level of competitiveness of OpenStack with other open CMPs.

The overall difficulties of the project cause many vendors to revise their OpenStack-related strategies. Often, suppliers, as well as OpenStack users, say completely different things about OpenStack openly and in private. This divergence between public and private opinions stems from users' desire to associate themselves with a project called the future of open cloud systems. However, their reservations about the ability of the project to create a stable product within a commercially significant time interval are very serious and, therefore, now many suppliers are reluctant to give significant technical resources to the project.

The desire of vendors to gain a competitive advantage in the OpenStack ecosystem leads to the fact that they are reluctant to give operational and technological capabilities to the open source kernel. For example, this kernel does not have an installer or upgrade or upgrade mechanisms, since both of these critical features are left for commercial distributions. Updates require reinstallation and a full upgrade of workloads. While the project is in its early stages of development, new versions can be radically different, thereby disrupting compatibility and requiring long periods of preparation and testing prior to the transition. Many early users of OpenStack still use older versions.

Service providers are interested in differentiated functions, and suppliers are interested in differentiated private distributions, or even selling additional software for OpenStack (including commercial CMPs operating at higher levels of the CMP software complex). They invest exactly as much in OpenStack for its success, and at the same time reserve as much resources as they need to ensure their own success.

Recommendations:

• If you are considering using OpenStack, perform a thorough technical assessment, determine whether you have the technical skills to deploy and operate OpenStack, and calculate the total cost of ownership of the solution because open source is not always cheaper.
• Consider licensed CMPs, as well as free and commercial distributions of OpenStack and other open CMPs, in particular Eucalyptus (which has received an official AWS API license from Amazon, and is relatively easy to deploy) and CloudStack (which is usually chosen by service providers and anyone who wants to use AWS-compatible cloud systems on a full scale).

Chief information officers may confuse the adoption of an API with the adoption of a CMP that supports this API, which forces them to narrow their technology platform to their own detriment.

The source API of the OpenStack platform was the AWS API, but it is in the process of migrating to its own native API, and continued support for the AWS API is a matter of controversy in the community. AWS competitors, particularly Rackspace, are concerned that this could strengthen the AWS ecosystem, so AWS compatibility may be removed from the main project over time. Currently, the native OpenStack API has only modest support for third-party tools.

since cloud APIs are relatively simple, it remains only to create a connector to translate one API into another. Regardless of whether you believe that OpenStack will become the most popular API over time, you do not need to use OpenStack CMP to use the OpenStack API; to do this, you can simply use the connector.

Recommendations:

• Don't limit your options. Instead of writing directly in the cloud IaaS API, use a third-party multi-cloud library such as libcloud or jclouds. Consider using a cloud management tool that supports more than one cloud, such as enStratus, RightScale, or ServiceMesh. Remember that the capabilities of cloud service providers can vary greatly, so anything that works with multiple clouds will be able to provide only inter-cloud functionality with the "lowest common denominator."
• Choose a CMP that supports multiple cloud API connectors, allowing you to quickly connect libraries and tools that work with other APIs.

OpenStack advocates argue that it will eventually displace full-featured CMP in commercial enterprises, especially VMware vCloud Suite, but it currently only implements basic CMP capabilities.

Supporters of OpenStack, particularly Rackspace and HP, argued that OpenStack would eventually become the preferred CMP for private clouds, including commercial enterprises, as well as service provider clouds. However, OpenStack's modest feature set makes it unlikely that it will displace more fully functional solutions. OpenStack is basically a resource administrator and can be at this level it will become a generally accepted component, but CMP functionality, and value, exist mainly above this level.

Over time, OpenStack can add components that provide higher-level functionality, but they probably will not have the same set of features as its licensed counterparts, as is typical for open source solutions in the rest of the IT management tool market. Each company associated with OpenStack services or commercial distributions has different strategies for entering the market, although in some cases, they may want to offer fully functional commercial CMPs that use OpenStack at the resource administrator level.

OpenStack solutions are unlikely to displace VMware vCloud Suite or Microsoft System Center, or competitive products such as BMC Cloud Lifecycle Management, within the next five years. Theoretically, OpenStack solutions do not depend on the hypervisor, but in reality they only support KVM or Xen technologies. Its support for VMware vSphere and Microsoft Hyper-V will have no functional value for enterprise deployment. It is also unlikely that OpenStack will displace vSphere or Hyper-V, in favor of Xen or KVM, in commercial organizations.

The OpenStack Alliance faces a very difficult battle with VMware, which already has a significant base of installed equipment, access to the market and a set of products whose scope is far beyond resource management. Microsoft is also rapidly increasing its market share, thanks to its strong position in the corporate sector and the fact that MicroSoft often acts as a strategic supplier. Customers want to be able to incorporate hybrid cloud features into their internal virtual environments, and typically turn to their old vendors for solutions.

Of course, many for-profit companies are looking for more cost-effective alternatives to VMware for hypervisor and CMP, and that many of them are interested in the potential cost reduction opportunities offered by open solutions. However, at the moment, such companies are much better off adding Microsoft as a second supplier than making an open decision.

In our opinion, in the next five years, most commercial companies will accept CMPs that integrate with the functions they already have, or that offer an easier to implement and more developed solution than OpenStack. This does not mean that OpenStack will not be successful over time, but it is unlikely to be the main corporate solution in the near future.

Recommendations:

• Do not make plans for the future development of your data center, which should be based on OpenStack. OpenStack can be considered a promising project, along with a large number of suppliers who have become victims of its noisy marketing campaign, but there are absolutely no guarantees in its future success. Instead of counting the number of companies joining the project, it is better to judge the progress of OpenStack by its ability to release future releases in a timely manner, along with planned features and a minimum number of fixes, and achieve universal recognition.
• In order to maximize deployment flexibility and interoperability, which over time will eliminate the use of multi-vendor services, try to choose CMP solutions that provide logical independence of service levels (access control, service management, service optimization, resource management and base resources) from each other.

OpenStack can eventually become a reliable open core of a large and successful ecosystem of commercial offerings.

There are a number of reasons why OpenStack can be considered a promising project, namely:

• The technical specifications and implementation are reliable and designed to deploy truly full-scale cloud systems.
• The breadth of the vendor community and ease of implementation quickly made OpenStack one of the top four cloud infrastructure ecosystems (along with Amazon, VMware, and Microsoft.)
• The basic platform provided by OpenStack allows new developments to pass the incubation period. In addition, the open character of the source code allows decentralized and collaborative introduction of new solutions.

We believe that there are three main ways OpenStack can succeed in the long term:

• Open source will eventually turn into a successful solution for those looking for basic, cheap CMP.
• Successful commercial products and services will be included in the open source kernel, in addition to the private extensions and capabilities added by vendor companies.
• IaaS cloud service providers will use the open source core, combined with proprietary capabilities or sourced from commercial providers, to strengthen their offerings on a large scale.

OpenStack is in the process of being transferred from Rackspace to the nonprofit OpenStack Foundation. Once this organization takes over the reins, it will have to answer questions critical to the development of the project, including defining the essence of OpenStack. For example, what can be called "OpenStack" and in what context.

Potentially, this could solve some of the interoperability and fragmentation problems, as well as strengthen the overall management and administration of the project, and will be critical for turning OpenStack from a highly specialized evolving solution into something suitable for common use.

Recommendations:

If you accept OpenStack:

• Join a community that needs stronger operator and user representation.
• In
  turn, offer your own fixes and improvements for inclusion in open source, if your organization allows it.

• Original article

2011

Cisco and Canonical join the project

In early 2011, it became known that Cisco and Canonical were joining the projects. The idea of ​ ​ cooperation with the OpenStack community at Cisco belongs to the new vice president and CTO of the company, Lew Tucker, who, before his employment at Cisco a few months before, held the same post at Sun Microsystems. As part of the OpenStack project, Cisco will deal with cloud network infrastructure technologies: "In order for cloud computing to fully cope with the applications of tomorrow, in our opinion, the network should be considered an infrastructure service, along with computing and storage services," writes Lew Tucker in his personal blog. He also notes that Cisco's participation will not be limited only to design and development of specifications - the company's engineers will also develop code as part of OpenStack projects.

Canonical's participation in the OpenStack project turned out to be somewhat unexpected, since so far this company has supported another open source project with similar goals - Eucalyptus. Just a day earlier, on February 2, Canonical, together with Eucalyptus and Dell, announced a solution to create private cloud systems. OpenStack is a direct competitor to Eucalyptus, and at one time, it was the controversial commercial model and architectural limitations of Eucalyptus that prompted NASA to create its own cloud infrastructure, which was the basis of the OpenStack project.

At the same time, the current accession to OpenStack is only a trial step and does not imply the rejection of partnership with Eucalyptus:

'Our long-term goal in the case of OpenStack is to make Ubuntu the best OS for'clouds' in this stack, both at the infrastructure level and as a guest system, 'writes Nick Barcet, head of cloud computing at Canonical. - In release 11.04 (Nattyy Narwhal), OpenStack 2011.1 (Bexar) will be included as a technology demo, and Canonica will not fully support this solution yet. We want to give our users the opportunity to test this solution and leave their feedback before we start positioning it as an environment ready for industrial operation '.

The OpenStack project at this time has the greatest appeal for developers who want to experiment with cloud computing. At the same time, the main difference between OpenStack and Eucalyptus, another well-known open source development for deploying private clouds, is absolute openness. If Eucalyptus provides full functionality only to customers of the enterprise version, then the functionality of the free version of OpenStack is not limited by anything. However, if pilot implementations are successful, companies begin to show interest in commercial support and maintenance services.

The announcement of paid services shows that OpenStack is not a completely non-commercial project. Cloud Builders' commercial service will include training and certification services, integration and support for OpenStack implementations, according to a company press release. However, the site is currently missing any pricing information for Cloud Builders commercial services.

Since OpenStack is an open source project, formally nothing prevents its other participants from starting to provide similar services, becoming competitors to Rackspace. Until recently, such a competitor could theoretically be the Anso Labs development group, which worked on OpenStack as part of orders from the NASA aerospace agency. However, after Rackspace acquired Anso Labs, fully concentrating the key development roster at home, the emergence of full competition for the Cloud Builders project soon became unlikely.

OpenStack Cactus Release

April 2011: Cactus is the third release of the project in less than a year since launch. The new release uses an imaging service introduced in a previous version called Glance. According to OpenStack Chairman Jonathan Bryce, in the new release, Glance has reached maturity and is fully ready for industrial use. Accordingly, the Glance service received a new, more respectable name: OpenStack Image Service.

Starting with Cactus, OpenStack will ship with a new API that will allow you to access the open cloud stack from a wide range of third-party applications. "The new API allows you to create extensions," says Bryce. "And in the API namespace, a special area is allocated for functions that are not related to the basic functionality of OpenStack software." In particular, the extensible API will allow you to program complex network functions or integrate OpenStack with equipment that is not provided for basic delivery.

Bryce expressed hope that the new API will promote innovation and allow more people to benefit from OpenStack. In addition, additional functions, originally written as modules, will be able to go to the main project code in the future.

The Cactus release also includes an advanced authentication system for cloud storage systems: "This is a completely rewritten distributed authentication system for the needs of the project, which is now ready for industrial use," Bryce said.

In general, the development and standardization of the authentication system is one of the priority areas for the development of the project, and this course will continue in further releases, Bryce added.

The next release of OpenStack is expected in July 2011 - it will be codenamed Diablo.

Riverbed Technology joins the project

Riverbed Technology, which specializes in IT performance solutions, announced in November 2011 that it was joining the OpenStack community, a global community of developers collaborating to build an open cloud architecture and open source cloud operating system. As organizations move their data and applications to the cloud, one of the main challenges in providing high application performance is the ability to network over a WAN network. Joining the OpenStack community, Riverbed is working to support the OpenStack platform in its WAN optimization solutions to create efficient cloud environments.

Providing support for OpenStack in Riverbed solutions for WAN optimization continues the company's existing collaboration with the OpenStack project, in which Riverbed provided support for this platform in its virtual application delivery controllers (vADCs). Riverbed WAN optimization solutions and virtual application delivery controllers provide a technology solution that enables organizations to improve the performance of their IT infrastructure in a public, private, and hybrid cloud environment.

"Mirantis" launches the site www.openstack.ru

In December 2011, a new platform for communication and interaction between Russian developers working with advanced OpenStack technology will start - www.openstack.ru. The community, which will bring together IT specialists, was created under the patronage of Mirantis and is designed to become a convenient platform for sharing experience and knowledge related to OpenStack, Cloud Computing, Private Cloud and many others.

OpenStack technology is one of the new solutions, the popularity and demand of which is rapidly gaining momentum. It includes a series of interconnected projects that provide the development of numerous components of an infrastructure solution for the "cloud." OpenStack is used by large companies, service providers, VARS, representatives of small and medium-sized businesses, researchers, as well as worldwide data centers. The OpenStack code is freely distributed: anyone can run the software, add additional settings, or return it to its original state.

Since Russian developers, including Mirantis specialists, are actively involved in the development of this area, one of the key tasks was to create a convenient and informative Russian-speaking platform. Its goal is not only the placement of various information on OpenStack technology, but, above all, the discussion and exchange of experience that developers have accumulated at the world level regarding the creation and use of OpenStack. Thanks to the web portal www.openstack.ru Russian specialists have the opportunity to share expertise, express opinions and receive data on "cloud" technologies online.

In addition to describing OpenStack technology and blogs that discuss the latest changes and news, the OpenStack website provides visitors with the opportunity to get acquainted with cloud events - for this, there is an event calendar on the site. It includes all conferences, seminars and meetings that are devoted to OpenStack technologies and are held in various countries of the world. In addition, there is a section "Vacancies" on the site, where developers can find an up-to-date list of open positions.

Boris Rensky, co-founder and vice president of Mirantis, comments on the appearance of the www.openstack.ru website: "OpenStack is one of the most advanced modern IT technologies, which is very quickly gaining popularity. Since OpenStack accumulates expertise from developers around the world, the exchange of experience comes first in the development of the product. Western specialists have a popular and popular resource - www.openstack.org, which helps both developers and companies - users of OpenStack technology to share up-to-date information and get answers to current questions. In my opinion, the contribution of Russian specialists to the creation of OpenStack is significant. Organizing a professional community dedicated to this technology is a response to a rapidly growing market and a platform that, I hope, will contribute to both the development of the solution and the professional growth of specialists in this area. "

2010: Project Start

The OpenStack cloud infrastructure project was opened on July 19, 2010 at the initiative of hosting provider Rackspace and NASA to counter Amazon's offensive on the cloud front. Currently, OpenStack includes all the services necessary for the cloud platform, including basic tools for computing, storage and networking, as well as additional functionality for launching common software technologies such as Hadoop and containers.