Anil Uberoi

Subscribe to Anil Uberoi: eMailAlertsEmail Alerts
Get Anil Uberoi: homepageHomepage mobileMobile rssRSS facebookFacebook twitterTwitter linkedinLinkedIn


Related Topics: MultiTouch Developer Journal

Multi-Touch: Article

The Cure for What Ails System Administrators

Linux Data Center Automation

As data centers become more complex, CIOs are under pressure to find ways to streamline/standardize their IT operations, create error-free, repeatable processes, and reduce the data center operating costs – all at the same time. In addition, as more and more business-critical applications and services are deployed on Linux, enterprises must have IT operational resiliency and the agility to respond quickly to changes in business needs. To accomplish these objectives, data center operators must implement a low-touch, dynamic management infrastructure that can monitor system states, analyze conditions, and perform remedial actions based on operational policies in near real-time. Some of the more contemporary data center automation (DCA) solutions are designed to address these challenges.

Let’s consider some practical applications. If a specific application server or a group of servers gets overloaded, additional capacity might be required. Although a traditional management tool would alert IT staff to this condition, the staff would still have to invest time and manpower in keeping that part of the data center up and running. In contrast, with a modern DCA solution it’s possible to both monitor the load in a group of servers and bring on additional capacity instantly by repurposing some of the servers in the existing server pool to different roles (for example, changing a reporting server to an application server). Another instance where the right DCA solution can alleviate some of the stress on sysadmins is in the case of a critical system failure (a nightmare for an IT team without proper tools). Advanced DCA solutions can monitor for a critical system failure and then, if a system does go down, remediate by quickly migrating the entire software stack – OS and applications – to replacement hardware in minutes.

In essence, an advanced DCA solution ought to automate a lot of these mundane repetitive functions that cost sysadmins an enormous amount of time, energy, and money, even assuming they have the expertise to perform such functions error-free. They need to allow for greater business agility (with minimum downtime by preventing quicker responses to system errors) at the lowest possible cost (as mentioned above, data center administration costs, specifically for data centers running on Linux, can be astronomical). The more data center management functions that can be automated, the less system administrators need to touch, thereby reducing the potential for errors. Clearly, for large complex enterprise data centers with many moving parts, the right DCA solution is important to maintaining a high level of efficiency and efficacy in IT operations, which, in turn, leads to smoother business processes.

The next issue to consider is the type of solution appropriate for a given enterprise data center. Are all DCA solutions created equal? For Linux-based environments, DCA choices range from Linux-specific point solutions that solve a portion of the DCA challenges to multi-platform DCA solutions where Linux is only one of many operating systems covered. Current DCA offerings can be grouped into three basic categories: the Big 4 frameworks, independent DCA vendors, and Linux-specific point solutions.

Although Linux has become a popular data center platform, many DCA vendors haven’t addressed automating a pure Linux environment. The Big 4 frameworks (IBM Tivoli, HP OpenView, BMC Patrol, and CA Unicenter), as cross-platform solutions, try to manage anything and everything in the data center and, as a result, become, in theory, jacks-of-all-trades-and-masters-of-none, which is problematic for enterprises that require a “master of Linux” solution. While good solutions for multi-OS data centers or those that don’t run business-critical applications on Linux, they become hard to use for enterprise data applications and services that rely on Linux; inefficient toolsets often adapt poorly to Linux packaging, some even offer different tools and approaches for each Linux distribution. They generally use heavy scripting and often lack flexibility while requiring costly implementation, integration, and customization efforts (generally in the form of professional services provided by the vendor). This all adds up to significantly higher TCO, not to mention the opportunity cost of a long time-to-implementation often of several months.

Independent DCA vendors can be divided into two separate groups: those that tune their DCA offerings to a specific platform (e.g., proprietary products like Sun N1) or those that try to be cross-platform like the Big 4 (i.e., BladeLogic, HP Opsware) with minimal Linux support included and often as an afterthought. The platform-specific players optimize for their own platforms and don’t translate well to Linux; for example, Egenera and VMware work best when managing systems within their own domain. The other independents have grown their capabilities incrementally without reaching true DCA. Like the Big 4, they focus on breadth, not depth; for a Linux-based data center. This could become a fatal flaw. Linux distribution vendors offer point solutions for their own flavor of Linux distro; for enterprises running on different Linux distros (one of the inherent complications of running a data center on Linux), this only solves a small part of the problem.
 

More Stories By Anil Uberoi

Anil Uberoi, CMO, comes to Levanta from BackFlip Software, where he was president. He has extensive marketing experience in enterprises with a variety of foci. Prior to his work at BackFlip, he was senior VP of marketing and business development at XACCT until its 2004 acquisition by Amdocs, where he then became VP of global marketing. Earlier, Anil led marketing for the networking products group (including the network management platform) at Sun Microsystems.

Comments (0)

Share your thoughts on this story.

Add your comment
You must be signed in to add a comment. Sign-in | Register

In accordance with our Comment Policy, we encourage comments that are on topic, relevant and to-the-point. We will remove comments that include profanity, personal attacks, racial slurs, threats of violence, or other inappropriate material that violates our Terms and Conditions, and will block users who make repeated violations. We ask all readers to expect diversity of opinion and to treat one another with dignity and respect.