OpenStack Building Enterprise Private cloud to solve five major problems

OpenStack has become a trend, but the release of OpenStack is not perfect, enterprises to build a private cloud must be fully aware of the shortcomings of the release OpenStack, and seek professional OpenStack provider help and cooperation in order to avoid weaknesses, Truly leverage the benefits of OpenStack and build a private cloud that maximizes your competitive edge.

How does OpenStack work well in the enterprise? What other issues need to be addressed? How can OpenStack be used in the enterprise? Developers think it is the use of posture, users think to be stable and reliable, not old downtime; The boss thinks that the development and operation of several cattle x can be done.

In fact, the problem of OpenStack in business, mainly in the following five aspects: stability, integrity, high availability, ease of use, dual live and disaster tolerance.

First of all, stability. A good product, performance is not the first factor, the stability of the enterprise is the most important.

A.openstack is still far from sufficient in scalability and stability, and needs to be carefully polished.

From dozens of to thousands of Taiwan, or even tens of thousands of units, whether the stability of the same as always can not be a problem? As the scale expands, the overall structure needs to do a sufficient job in terms of stability.

For example, you need to design multiple Nova APIs and multiple mirrors, load balancing and node high availability, and database concurrency responses.

In addition, in the community by the most up-to-the-top upgrade problem--nova,swift,cinder and neutron use their own database storage configuration information, to upgrade the database schema to modify multiple, do not have a hot upgrade (after the H version of the upgrade problem is improved).

Again, when a business deploys network services (Neutron), it encounters a nightmare experience and has to rewrite the code of the network components to meet the demands of large-scale applications.

B.openstack lacks integrity.

A mature cloud platform should provide computing, storage, networking, security, database, big Data, middleware, DEVOPS, monitoring operations and many other cloud products. OpenStack can only provide computing, storage, network three kinds of cloud products, if enterprise customers need information security protection products, it must be self-service information security platform, integration of third-party products. Like Big Data analytics, where Hadoop clusters can be deployed quickly through Sahara, how do you get through the account, security, management, and operational monitoring systems between OpenStack and Hadoop?

The high availability of c.openstack at the virtual machine level is not good enough.

There is no official statement that OpenStack supports high availability at the virtual machine level, which is presented in the Folsom version, but is later discarded.

OpenStack currently has an incubation project evacuate, which is used to provide high availability support for OpenStack at the virtual machine level. Evacuate currently can only be initiated manually by the administrator, evacuate does not consider the deployment properties of the VM, causing the resource scheduling policy to fail. The change in hostname causes the Nova-compute to mistakenly delete all virtual machines during the restart process, and the problem arises mainly because of the evacuate cleanup mechanism. This bug has been fixed in the L version.

The ease of use of d.openstack is not good enough.

With fuel, OpenStack can be installed quickly, but many configuration operations also require a command line, and a key delivery distance from the automated deployment. Again, for example, the more extensive ceph distributed storage system used by OpenStack has not yet implemented interface-based operations and configurations. In addition, OpenStack lacks a common base version.

The use of OpenStack will not be locked by the vendor, but OpenStack can download more than 20 vendor-specific versions, and the customer's choice is important.

E. Double-live and disaster-tolerant issues.

Large enterprises on business continuity requirements, key core business has the same city double live and remote disaster recovery needs. Double live in the same city refers to the user's critical business system at the same time in the city's two data centers running, while providing services to users, when a data center application system problems, there is another data center application to continue.

Disaster recovery, as the name implies in different regions, the construction of a set or more than one set of identical applications or databases, play a role in the immediate takeover after the disaster. We see that OpenStack, while also having a single-site (Smaug+cinder) and cross-site (smaug+swift) Backup and recovery solution, is far from true business-to-life and offsite disaster.

such as tricircle implementation of the cross-datacenter cascade, or need to cinder rely on the storage back-end their own ability to do disaster preparedness, tricircle itself as a forwarding relay, for users to find the right operation of the site, which itself can not achieve across the data center disaster-tolerant function, This is different from VMware's SRM.

As we can see, there is still a gap between OpenStack and VMware in terms of feature support and specifics, and there is a need for continuous improvement to do better. But OpenStack, as an open-source management framework, is designed to be good. With the use and development of OpenStack in the enterprise, it will promote and accelerate its maturity.

Finally, operation and maintenance automation, in the large-scale cloud operation and maintenance scenarios, the need to be high-repetition, based on intelligent decision-making monitoring data trigger, to achieve unattended automatic operation of the operational capacity, which is still to be explored by OpenStack.

Conclusion

The process of building a private cloud is not an overnight one, and it will be an ongoing process of engagement that needs to be constantly explored in practice. In the next 10 years, will be the Cloud 2.0 era, the enterprise is the main representative of cloud. Looking to the future, embrace open source. In China, in Asia, and globally, OpenStack has become a trend, and we have reason to believe that OpenStack will be the first choice for enterprises to go to V in the next three years.

OpenStack Building Enterprise Private cloud to solve five major problems