Data Centre Design: Improved Performance & Efficiency

Many companies are redesigning their data centres in order to improve their performance and increase efficiencies, an exclusive survey of IT executives shows.

The IT manager at a large enterprise is frustrated. "We've done the simplification thing over the last few years and we still have overlapping systems." The partially-centralized IT infrastructure he refers to and which is common, is riddled with inefficiency and inconsistency that hinders business process performance and increases costs. "We are supposed to be the centralized data processing for our organization and manage networks, but we also have departments such as human services, which have their own network staff, and other departments that have their own server/desktop teams. We have departments that just have desktop people and others that don't have any IT people. This leads to very inefficient, inconsistent operations," the IT manager explains.

This IT manager isn't alone. After a decade or more of decentralized, distributed information systems, many organizations are realizing that decentralization too often costs them efficiency and reliability. As a result, they've committed to moving their distributed, decentralized systems back into a central data centre, where they can take advantage of the latest advances in technology, such as WAN acceleration and virtualization, in order to improve data production and boost enterprise-wide IT performance.

Driving interest in central data centres are a number of IT trends, including data centre consolidation, support for server and desktop virtualization, cloud computing, disaster recovery and business continuity (DR/BC), and Service Oriented Architecture (SOA). Using these strategies and others, organizations are attempting to deliver adequate application performance to users with the fewest possible problems and at the lowest possible cost. Unfortunately, in a survey conducted (Q3 2008) of more than 100 IT executives on the subject of improving data centre performance, IDG Research Services found that data centre costs are escalating rapidly and many are running out of both power and space. The survey suggests IT managers must redesign their data centres to consume less space, less power and less monetary investment while still meeting performance and reliability requirements.

Data Centre Performance Challenges

Although it sounds straightforward, improving performance and lowering costs through network simplification is actually quite challenging. After all, just because managers declare a standard doesn't mean that their staff will follow it. A number of factors may hinder organizations' attempts to improve data centre performance. These include:

Growth: Even basic growth causes a physical space crunch in the data centre and increases the consumption of energy, which increasingly is constrained by power utilities. More networking, more servers and more storage continue to eat up costly floor space and consume power.

Complexity: More devices/protocols in the data centre cause more complexity, increasing the need for interconnectivity and interoperability at all levels, not to mention the need for more/better management and more/different training.

Software: More systems and applications increase the management problem, requiring new tools that can provide greater visibility across protocols, networks, systems, devices and applications.

Security: Increased complexity results in security gaps due to inconcsistent policies and the inability to apply, enforce and integrate policies across different devices, networks and systems.

Disaster Recovery/Business Continuity: In the event of disaster, data centres demand protection that requires consistent, reliable replication. This consumes costly bandwidth and must be managed efficiently to ensure reliability and performance within reasonable cost constraints.

Virtualization: Virtualization increases utilization of network resources but adds complexity, which places greater demands on QoS and imposes low latency performance requirements on the network.

Of these challenges, complexity is perhaps the most significant, as growing complexity in the data centre slows application performance, delays the deployment of new business-enabling applications and hinders innovation. In IDG's data centre survey, more than half of respondents (59 percent) indicated that the level of complexity in their company's data centre network has increased compared to 12 months ago, suggesting that the obstacles to simplifying data centres are mounting, pushing many toward a breaking point that's characterized by inefficient use of infrastructure resources and higher operational costs, as well as increased security risks or downtime due to human error and network failures.

Of course, complexity isn't just a problem. Often, it's also a symptom that's been caused by other system flaws. For example, 77 percent of IDG survey respondents said that software is the primary factor contributing to the complexity in their data centre networks. Meanwhile, nearly half of respondents (49 percent) attributed the complexity to hardware and 43 percent say it is due to operating systems and devices.

Technology Complexity

IT environments are growing more complex. - so they may suffer from technology sprawl (defined as uncontrolled proliferation of servers, systems and devices). Different operating systems and management tools make it difficult to integrate and create interoperation. Technology sprawl reduces data centre performance and efficiency, which raises costs. Virtualization, a priority among survey respondents, reduces data centre technology sprawl, but brings new complications.

Being a relatively new technology, virtualization is difficult to manage correctly due to immature management tools and evolving best practices. Management tools, certainly, will develop in the future. In the meantime, the majority of organizations (68 percent) report investing today in server virtualization, driven mainly by cost advantages.

Organizations rightly view applications as a critical asset. However, software also poses challenges to improving data centre performance, as software in the data centre extends from applications to network operating systems. In fact, there are operating system-derived interoperability and integration challenges among the diverse network elements in the data centre. As a result, software management continues to be problematic.

Security may be even harder. A steady stream of headlines reporting IT security breaches that have compromised the privacy of millions reminds IT managers of high-profile IT risks. These concerns were reflected in the IDG Research study, in which 43 percent of respondents reported concerns with security, especially risks introduced by new technology like virtualization. Fueling these concerns is the growing demand for network access to data and applications, including remote and mobile access, which also leaves organizations vulnerable to human error.

Balancing Cost and Performance

The IDG Research survey found that 74 percent of respondents are ready to invest in information technology to improve data centre performance-especially if the investment would lead to lower operational costs or improve service delivery (66 percent). Other issues driving data centre IT investments at these companies are the need to increase productivity (58 percent) and compliance requirements (53percent)

"The primary concern we are hearing today from IT managers about the data centre is high performance with cost avoidance," says Bobby Guhasarkar, senior manager, Juniper Networks. "Even if something improves data centre performance, they also need it to reduce operational costs."

Reducing costs is not an easy goal to reach, however-especially in more complex data centre network environments.

Cost cutting is an unavoidable necessity, suggests the IDG survey data, and organizations are doing whatever it takes to facilitate savings. For example, the survey found that a need to cut costs is driving many respondents (63 percent) to improve data centre performance by redesigning their data centre network. For these organizations, cost cutting isn't just about saving money. It's also about supporting new tools with which to make money. For instance, among the leading economic drivers behind initiatives to boost data centre performance, according to IDG data, are the need to make room to support new applications (47 percent) and the need for 24/7 availability (46 percent).

In addition to lowering costs by reducing the amount of equipment, space and power they consume, companies are turning to the following revenue-generating objectives to motivate data centre improvement initiatives:

Improving service delivery: Companies want to lower latency with better and more scalable security and advanced network routing between data centres and across the WAN.

Increasing productivity: Companies want to gain increased efficiency through faster and easier provisioning of applications and network resources.

Achieving compliance: Companies see data centres as making it easier to meet regulatory mandates and internal corporate governance policies. Still, it all comes back to cost. In fact, the IDG Research study found that the primary investment driver for respondents (62 percent) was the need to lower costs. Next, they said they wanted to improve service delivery (60 percent), support new applications (52 percent), improve security (32 percent), maintain compliance (28 percent) and improve scalability (25 percent).

Data Centre Technology Concerns

IT managers have concerns about sustaining the performance of their data centres in the face of growing workloads. These concerns, according to IDG Research survey respondents, revolve around:

Rapid problem identification and resolution: IT managers need to achieve management and event visibility across the entire data centre and network infrastructure.

High availability: IT managers need accessibility at all levels-application, system, platform, and network.

Complexity-related human error: IT managers need a consistent, easy-to-learn, easy-to-use provisioning interface with which to access network and data centre technology.

The IDG Research survey identified several additional top technology interests that suggest the importance of network security and management capabilities to data centre performance. For example, respondents report they are deploying security/access management (58 percent), Web services (58 percent), security and vulnerability management (56 percent), and data centre management (47 percent).

Optimizing Data Centre Performance

Despite the challenges, it is possible to boost data centre performance while lowering total cost. Industry analysts and vendors identify the following steps an organization can take to optimize data centre performance:

Redesign the data centre to streamline processes

Collapse data centre network tiers and reduce the L2 complexity

Design networks to deliver appropriate QoS for each system, application and user community

Consolidate data centre security to ease and improve policy management

Leverage advanced routing capability to support performance and security requirements

Design with interoperability and integration in mind

Consolidate network management through a single, intuitive console

Integrate support for new systems and new types of applications in the future

 Include environmentally efficient data centre practices

Of the above solutions, Guhasarkar suggests that one of the most important is consolidating the overabundance of devices within your data centre. "Every application serially traverses all these devices, which adds latency to every transaction," he says. Guhasarkar adds that removing some devices by collapsing switching tiers in the data centre will reduce complexity and latency, as well as costs, as you'll be eliminating systems and devices that are no longer needed.

Although consolidating equipment is a good start, many organizations are attempting to simultaneously improve performance and reduce costs by making organizational and process-focused changes. "We are attacking the entire set of obstacles on several fronts," explains the manager of a mass merchandiser. "We have expanded our change control process to help minimize human errors that occur during system changes. We implemented policies and procedures for a variety of tasks, including system builds, network changes and application updates. The objective was to sustain and gain consistency; we brought in consultants to monitor our systems and engage more proactively with the issues before the user community was impacted. We've added redundancy in our systems and networks wherever physically and fiscally possible."

Improved Performance

Although optimizing one's data centre requires time and effort, the return on investment can prove invaluable, as improved data centre performance impacts not only IT, but the entire organization. Improved data centre performance also enables better compliance with regulatory mandates and internal corporate governance policies. By consolidating and redesigning the data centre network infrastructure, organizations can reap considerable performance gains that deliver operational simplicity, business agility and greatly improved efficiencies.

At the heart of this redesign will be a single efficient, flexible, low-latency network.

About Juniper Networks

Juniper Networks Inc. is the leader in high-performance networking. Juniper offers a high-performance network infrastructure that creates a responsive and trusted environment for accelerating the deployment of services and applications over a single network. This fuels high-performance businesses.

www.juniper.net/uk/choice