First, a common measurement for data center energy efficiency is PUE or power usage effectiveness. PUE is a standard ratio that looks at the total energy supplied to a data center and divides by the amount of energy that actually reaches the data center equipment.  Mr. Niccolai reports, “It reveals how much energy is expended on cooling systems and other non-compute functions.”  Here is one example from the article: a PUE of 2.0 means that for every 2 watts supplied to the data center, only 1 watt reaches the computing equipment.

A survey by Digital Realty Trust looked at 300 IT decision makers at large corporations. The average PUE was 2.9.  An ideal score is close to 1.0.  Although this number caught many experts by surprise because past surveys had seen slightly lower scores, Digital Realty Trust felt 2.9 is probably closer to the “norm” and the figures underline the balancing act between service availability and energy efficiency, “IT is not easy. Designing data center operations around an organization’s infrastructure and operations is not a simple task.”

Mr. Niccolai reports the survey did find many efforts underway to improve data center energy efficiency. In fact, four out of five respondents said they are taking steps to keep hot exhaust air from servers mixing with cold air used for cooling, known as “hot-aisle or cold-aisle containment.”  As well as 85% use some type of data center infrastructure management software that can help to improve energy efficiency.

Mr. Niccolai shared these other survey findings:

• 66% of the survey takers have built or acquired a new data center in the past two years.
• 25% of survey takers reported operating six or more data centers.
• The average power density is 8.5kW per rack and the average IT load is 2.6 mW both up from last year’s survey.

A Gartner survey found that 40% of virtualization projects were undertaken without the security team’s involvement. In a legacy world, a security threat to one server may only put one workload at risk, but in a virtualized data center the hypervisor has oversight over all workloads running on a physical server putting them all at risk reports Mr. Brodkin. Experts recommend the following virtualization best practices to follow when beginning a project:

• Always involve the information security team early in the initial architecture and planning stages.
• Establish policies regarding the consolidation of workloads of different trust levels. Also when evaluating new security management tools Gartner recommends looking for those that span physical and virtual with the same management, policy and reporting framework.
• Once installed ensure patching, removing old users accounts, etc. continue to be up to date.

In terms of storage, Mr. Brodkin reports that “virtualizing servers without adapting physical storage systems to the unique needs of virtual machines is a kiss of death to any virtualization project”. Two virtualization best practices are thin provisioning and wide stripping technology. Thin provisioning pre-allocates a larger amount of storage to applications than is actually available to them, allowing applications to share a pool of storage and use only what they need, when they need it. Wide stripping technology helps to solve the increase in I/O patterns from virtualization by distributing the I/O load across many disks, rather than just a few.

How can cloud computing benefit the organization, what are the risks? In other words, what are the pros and cons of cloud computing. Tech Junkie provided a few straightforward facts to consider:

Pro: Cloud Computing allows IT to spend more time innovating and less time on maintenance since the cloud service provider. In fact, a recent survey in Forbes Magazine found businesses “relying on cloud computing strategies to free up dollars and time that can quickly be re-invested in product and service innovation.” In fact, the survey found 62% of respondents were able to invest more money back into their business due to cloud computing.

Pro: Cloud computing offers flexibility not only in the ability to scale up and down but also in terms of accessibility as employees are no longer restricted to the office.

Con: Cloud computing transfers some of the responsibility of protecting and encrypting your data to the cloud service provider which introduces risk.

Con: Cloud computing also introduces a new risk to availability. Instead of managing 100% hand-ons the availability for the organization, the cloud service provider will manage a portion of availability for the areas serviced.

Just like with any other new technology, cloud computing has its share of pros and cons. A thorough vetting of the cloud computing service provider in addition to a thoughtful and well-developed cloud computing strategy can help an organization mitigate the risks while reaping the rewards.

Let’s take a look at some common data center outages and the recovery checklists that correspond:

Power Outage- Data Center Recovery Checklist

• Step 1: Determine the extent of the outage.
• Step 2: Determine if staff needs to evacuate.
• Step 3: Assess potential damage to firm, ensure critical data is backed up and protected.
• Step 4: Contact senior management.
• Step 5: Contact utility company.
• Step 6: Identify cause of outage, launch remediation efforts.
• Step 7: Assess when data center operations can resume.
• Step 8: Contact senior management, send regular updates on progress.

Server Failure – Data Center Recovery Checklist

• Step 1: Determine extent of server outage, data loss and other potential outcomes.
• Step 2: Launch remediation efforts, e.g. check power supply, attempt server restart, run diagnostics.
• Step 3: Assess potential damage to firm, ensure that critical applications and data running on server(s) is backed up and protected.
• Step 4: Identify cause of server outage, continue remediation efforts.
• Step 5: Assess when normal server operations can resume.
• Step 6: Contact senior management, send regular updates on progress.

Fire – Data Center Recovery Checklist

• Step 1: Assess nature and extent of fire.
• Step 2: Use existing fire suppression equipment to extinguish fire, e.g. sprinklers, hand-held extinguishers. Note: if fire is severe move quickly to call 911 and/or evacuate staff.
• Step 3: Dial 911, advise of situation.
• Step 4: Evacuate building staff.
• Step 5: If possible and safe activate data center backup measures to protect current data.
• Step 6: Once fire is out, begin damage assessment.
• Step 7: Update senior management on status.

According to Fox Business, the Red Cross finds that 40% of small businesses do not reopen after a major natural disaster such as a flood, tornado or earthquake, but experts point out that even smaller incidents such as a fire or pipes bursting can threaten the financial future and security of a small business.

What does your disaster recovery plan look like? Creating a plan can take time, but it is worth every minute when an event occurs. Paul Mah of IT Business Edge provides some starting points specific for small businesses to consider when developing a disaster recovery plan:

• Back up work data. Digitized data is most likely the most “precious asset” of your business. Step one of a disaster recovery plan necessitates that a copy of this data be regularly kept. A location that is geographically separate from the business but easily accessible is best.
• Restoring server assets. Mr. Mah points out that “having a backup copy of data on hand is good, but of little immediate use if replacement servers required to make use of them cannot be deployed within a reasonable timeframe.” Infrastructure virtualization has allowed many small businesses to make copies of virtual machines with little effort that can typically be used in disaster recovery with sometimes only slight configuration needs.
• Working from home. Mr. Mah points out that not only does a small business need to account for the ability to restore data and infrastructure in a disaster event, but also how can its employees resume working if a disaster prevents them from being on site. Mr. Mah says to consider company laptops as well as a means to securely access work files.

Computer Weekly offers us advice from Gartner regarding data security issues to tackle, investigate and plan for before adopting cloud computing.

Breach notification and data residency. All data is not created equal and a blanket security policy should not be applied to all. In fact, Gartner recommends that data be categorized properly in importance as well as labeled with compliance or breach notification requirements.

Data management at rest. Know your cloud service provider’s data storage life cycle and security policies. Is multitenant storage used? What types of separation mechanisms are used? Does the key management strategy include strong identity and access management policies?

Data protection in motion. At a minimum to reduce data security issues in cloud computing, Gartner recommends the service provider supports secure communication protocols such as SSL/TLS for browser access or VPN-based connections for system access.

Encryption key management. Gartner recommends that the organization should always aim to manage the encryption keys. The Computer Weekly article further states that “if the keys are managed by the cloud service provider, then the organization should require hardware-based key management systems within a tightly defined and managed set of key management processes.”

Access controls. An organization should demand that the cloud provider offer adequate user access and administrative controls, stronger authentication alternatives such as two-factor authentication, management of access permissions and separation of administrative duties.

As David Linthicum points out for InfoWorld, “Many of the issues that cause IT organizations to push back on cloud computing – compliance, security, ownership and resiliency –are solvable problems, even though they require a bit of proactive planning. Share with us how you are planning to mitigate security issues in your cloud computing strategy?

With all the hype, it can be easy to lose sight of the need for careful planning and consideration when developing a big data strategy. Luckily, Gartner has provided a list of “essentials” for developing an effective strategy. Here is what they had to say in an interview with SearchCIO:

• Embrace the cloud. There will be a day when big data’s volume, velocity and variety will make more sense to store in the cloud versus in the company data base.
• Invest in skills. One of the biggest factors affecting the future of big data is a skills gap. In fact, in recent weeks, there have been many articles related to the future of big data especially in terms of job opportunities. The Boston Globe reports that a wide-range of companies are looking to expand their ability to make profitable connections within these new massive data streams.
• Pay attention to the role of data governance. Good data governance where data is cleaned, integrated, organized and labeled consistently is key to successful big data programs.
• Work hand in hand with business. Bridging the gap between business and IT has always been important but with business becoming more data-driven the importance increases exponentially.

The New York Times reports that “today, many experts predict that the next wave [growth and innovation] will be driven by technologies that fly under the banner of Big Data – data including Web pages, browsing habits, sensor signals, smartphone location trails and genomic information, combined with clever software to make sense of it all.” How could your organization’s future change with a solid big data strategy?

From air conditioning, capacity planning, location, infrastructure redundancy, power designs and fire-proofing, there can be much to think about in data center design. Today, we will concentrate on the general data center types and how they are defined with the help of Julius Neudorfer from Data Center Knowledge.

Standard Data Center. This design generally follows common industry standards and best practices in terms of layout, rows of cabinets and equipment according to Mr. Neudorfer. This type of design takes into account your organization’s size and future growth expectations to determine the facility size, redundancy and power need.

Build-to-Suit Data Center. Although the name implies “custom” this type of data center design actually offers only a limited degree of customizable options within the standard design.

Custom Design Data Center. As you might have guessed these data centers are designed to meet specific performance opportunities and needs. Typically only on the larger-scale end where dedicated and unique applications and hardware are required to operate.

An emerging trend in data center design is the shared data center or “colocation”. According to Forbes Magazine, data centers make up 2% of the energy consumed in the United States. While this is far lower, reports Drew Leonard, than other industries such as transportation, IT managers have an eye on energy efficiency. “In a colocation data center, multiple companies maintain their own IT infrastructure, efficiently sharing the facility’s space, power, security and network connectivity.”

First the dollars, Forbes Magazine contributor Louis Columbus outlines the top spending trends in cloud computing from Gartner:

• Global spending on public cloud services is expected to grow from $76.9 billion to $210 billion by 2016.

• Infrastructure as a Service (IaaS) is predicted by Gartner to be the fastest growing area of cloud computing.
• Compound annual grow rate (CAGR) for Platform as a Service (PaaS) is expected to reach 27.7% with Cloud Management and Security Services attaining 26.7% CARG by 2016.
• Global Software as a Service (SaaS) spending expected to grow from $13.5 billion in 2011 to $32.8 billion in 2016.
• The highest growth segments for the SaaS market will be office suites followed by digital content creation.
• According to Gartner, 59% of all new spending on cloud computing services will originate from North America with Western Europe accounting for 24% of all spending.

Gartner also sees cloud computing shaping security trends according to an article by Ellen Messmer of Network World.  By 2015, 10% of overall IT security enterprise capabilities are expected to be delivered in the cloud.  Analysts predict, “Increased adoption of cloud-based computing is expected to impact the way security is consumed as well as how key government agencies will prioritize security of public cloud infrastructure.” According to Ms. Messmer’s article today, the focus is on messaging, web security and remote vulnerability assessment, but Gartner expects more on the fronts of data-loss prevention and encryption as technologies mature.

One driving force to this security trend is the U.S. federal government’s FedRamp Program which accommodates external federal cloud computing services adoption and will allow for more cloud growth and use by federal agencies.   This growth could lead to public-cloud infrastructure becoming subject to “critical national infrastructure regulations” by 2016 according to Gartner.

Pick your migration strategy:

• Just data migration: typically the correct choice for Tier 1 and Tier 2 applications and will involve replication.
• Machine replication: typically best for Tier 1 and Tier 2 applications that can afford some downtime and involves stack migration.
• P2V migration: typically best for Tier 2 and Tier 3 applications that are not already virtualized.
• Disaster recovery: Mr. Robinson points out that some companies treat the migration as a disaster recovery scenario. They may choose to replicate the entire stack from point a to boint b and then click the failover button.

Next, Mr. Robinson recommends you understand the ‘gravity’ of your data. The data will be the weightiest part and he points out that since there is no easy way to shrink down data the weight of the app needs to be considered before migrating.

Finally, identify how your apps are connected. Are there other applications that access your data? Where will those applications live? Should or can they be migrated as well? How does latency affect the application? What is the sensitivity? Understanding how applications are tied together will help to mitigate post-migration surprises.