Gartner: Security Needs To Get Virtual

Security principles will remain fundamentally the same across virtualised data centres and private cloud environments, but how security services are provisioned and delivered will have to change, said Gartner on 8 November.

Most organisations will first virtualise servers in the data centre as a “stepping stone” before shifting to the private cloud, according to Thomas Bittman, a Gartner vice-president.

Integrated security

Security must be integrated from the outset and not “bolted on” later as organisations shift from virtual data centres to private clouds, said Bittman.

Virtual security will continue to be a hot IT topic, with Gartner predicting that 40 percent of security controls within enterprise data centres will be virtualised by 2015, up from 5 percent in 2010.

To secure private clouds, IT managers have to design a set of on-demand, elastic and programmable services that are tied to logical attributes to create adaptive trust zones, according to the analysts. Security must be an integral, but separately configurable part of the client environment, the analysts said.

Private clouds are likely to be deployed incrementally, not all at once, and will be carved out of existing data centres, said analysts. Security in the private cloud would need to exchange and share policies with other data centre security infrastructure, for both virtual or physical, wrote Gartner’s team of analysts. Security controls placed across either infrastructure would “intelligently cooperate” for workload inspection, said Gartner.

Physical hosts

Policies where the physical host is separated to isolate it, or “tied” to physical attributes, such as the server, Internet Protocol (IP) address, or Media Access Control (MAC) address, “break down” with private cloud computing, said Neil MacDonald, vice-president at Gartner and a Gartner Fellow. Policies approving or denying actions should not be based on physical components in the cloud because workloads and information are no longer restricted to specific devices, said the analysts.

Security should be delivered as a set of “on demand” services to protect workloads and information, and integrated into cloud provisioning and management processes, said Gartner analysts. Appropriate security policies that handle provisioning, moving, modifying, cloning, and retiring should be associated to server and desktop workloads, according to analysts.

Security policies should not be administered on a single virtual machine basis, said Gartner. Instead, security policies will be used to logically organise workloads into trusted zones, such as based on similar security requirements or trust levels, the analysts said. As policies are linked to groups of VMs and not physical servers, the zones adapt as individual machines and new workloads are provisioned and moved, said Gartner.

According to the research firm, private clouds will require security services that can separate workloads with different trust levels. By 2015, 70 percent of organisations will allow server workloads of different trust levels to share the same physical hardware, unless expressly prohibited by regulatory or compliance rules, estimated Gartner.

Just as there is a strict separation of duties and concerns between physical and virtualised infrastructure, IT operations and security must be separated on multiple levels in a cloud-based environment, said analysts.

Ideally, security policies designed to protect workloads should be federated to public cloud providers, said Gartner. While there are no currently established standards, analysts cited VMware’s vCloud API as an example.

Fahmida Y Rashid eWEEK USA 2014. Ziff Davis Enterprise Inc. All Rights Reserved.

Share
Published by
Fahmida Y Rashid eWEEK USA 2014. Ziff Davis Enterprise Inc. All Rights Reserved.

Recent Posts

US Finalises Billions In Awards To Samsung, Texas Instruments

US finalises $4.7bn award to Samsung Electronics, $1.6bn to Texas Instruments to boost domestic chip…

5 hours ago

OpenAI Starts Testing New ‘Reasoning’ AI Model

OpenAI begins safety testing of new model o3 that uses 'reasoning' process to ensure reliability…

5 hours ago

US ‘Adding Sophgo’ To Blacklist Over Link To Huawei AI Chip

US Commerce Department reportedly adding China's Sophgo to trade blacklist after TSMC-manufactured part found in…

6 hours ago

Amazon Workers Go On Strike Across US

Amazon staff in seven cities across US go on strike after company fails to negotiate,…

6 hours ago

Senators Ask Biden To Extend TikTok Ban Deadline

Two US senators ask president Joe Biden to delay TikTok ban by 90 days after…

7 hours ago

Journalism Group Calls On Apple To Remove AI Feature

Reporters Without Borders calls on Apple to remove AI notification summaries feature after it generates…

7 hours ago