OpenStack goal: Solve the bottleneck problem in implementation

With the advent of the new Year, the focus of OpenStack development will be shifted to address the bottleneck that prevents enterprises from implementing open source cloud computing platforms.

OpenStack developers have developed projects that mitigate the usability and scalability problems of the system, as well as provide OpenStack enterprise production instances, although these projects are still in the incubation period or have just been released recently.

One of the most prominent projects is a multiple project plan called Tripleo, and this project has an alias--openstack OpenStack. By using OpenStack to run its own facilities on bare metal as the foundation of the platform, this project can automate OpenStack installation, upgrades, and operational processes. OpenStack running on bare-metal will need support from another project within Tripleo, and this project is in the incubation period, which is ironic enough.

"It will also make it easier for operators to install on new machines," said Mark Atwood, an Open-source project manager at Hewlett-Packard. These carrier-centric and large-application-centric approaches will enable OpenStack to achieve greater success in the enterprise.

Another project, chaired by IBM and SoftLayer, is the benchmark facility, known as rally, which was recently opened to the public. Rally is an Open-source software that can be downloaded by cloud operators and developers who can use the software to develop a set of complex, highly repetitive tasks for benchmarking with OpenStack.

"Before rally, no formal development team had developed specific, highly repeatable benchmarks," said Michael Fork, chief architect of IBM's hosted private cloud computing, who helped design and develop rally.

Controversy over the use of reference architectures

Then there is the complexity of building OpenStack's underlying physical infrastructure, and a OpenStack Foundation board member believes that OpenStack can benefit from a standardized and repeatable reference architecture.

"Every OpenStack is like every snowflake," says Randy Bias, CEO and CTO of Cloudscaling. Cloudscaling is a contributor to a professional services company and OpenStack project that developed a 1.3 million dollar OpenStack deployment project for a large enterprise.

"Unless we have some reference structures that people can use, people don't know that OpenStack is ready," bias said. "Implementing OpenStack behind closed doors will only lead to the emergence of a custom island-as if you were developing your own Linux distribution." This also fundamentally limits the likelihood that the enterprise will achieve long-term success. ”

Others disagree that the reference structure is necessary.

"I don't think how much people need a good bundle Reference architecture, and even if there is such a need, there will always be companies to meet their needs," said Jonathan LaCour, DreamHost's cloud computing vice president. The company has developed an infrastructure based on OpenStack.

OpenStack Development waits for No

Analysts predict that 2014 of corporate IT spending will not shift to cloud computing, but some industry experts still believe that time is the most important for OpenStack development as an expected target for large enterprises.

"They had time until the end of the year before people began to question the viability of the platform," said David Linthicum, senior vice president of Cloud Marvell, a cloud computing consultancy in Boston. "I don't think OpenStack will fall into the realm of being unsuccessful because people will look for new directions and their businesses will benefit from it," he said. ”

The rift between the three main stakeholder groups in OpenStack must be repaired before continuing the OpenStack development, Bias said. These three groups are the developers, the end users of OpenStack cloud computing and the operators of these cloud computing.

The end users of OpenStack cloud computing want standardization and interoperability so they can move freely between different openstack cloud computing, Bias said. At the same time, cloud operators want to use their particular storage or network infrastructure, which makes their OpenStack versions look different from other versions of the OpenStack, thus severing contact with end-users who want to standardize.

Then there are the OpenStack developers who focus on innovation and add more functionality. In some cases, bias says, this will help improve stability and scalability, which represents a disconnect between the end users of cloud computing and cloud operators.

"A deterministic sign of OpenStack maturity is that most end users and cloud operators are pushing new features and implementing the product of work," bias said.