Server and Data Center Equipment Replacement Cycles


Replacing Servers & Data Center Equipment

Due to more conscious spending, tighter IT budgets, and/or more reliable hardware, companies are able to wait longer before having the need to upgrade servers and data center equipment. Traditionally, the server replacement cycle is every two to three years. Now, companies have been known to wait as long as six years before the need of any replacements.

Reasons to Upgrade Servers

There are no concrete rules in deciding when companies should upgrade servers. Decisions are typically based on server performance, rather than depreciation or when the hardware support and warranty contracts expire. This is the reason it is critical to continually monitor and take detailed reports of vital server metrics such as reliability, up-time, bandwidth consumption, and CPU utilization.

Building a Data Center

In addition to current server performance, it is important to calculate the future needs of the operation when building a data center. This is necessary in order to make smarter short-term upgrades and assists in anticipating a long-term server replacement strategy.

It is also important to consider all aspects of the IT infrastructure since each component is highly dependent on one another. Factors include projected processing and storage needs, future operating systems, databases, and application deployments. In doing so, a company can mitigate the risks of running new software on under powered hardware or old software on overpowered hardware.

“Reduce, Reuse, Recycle”

With the recent buzz surrounding green IT and efficiency, it can be beneficial to reuse old servers for less demanding processes. For instance, old production servers can be used as backup servers since the decrease in performance may be acceptable for intermittent use. Although this is a good way to cut costs, it is always advisable to analyze whether re-purposing hardware is more cost-effective than a server upgrade. Either way, strategy should be based on meeting the needs of the business.

Final thought:
The best approach in determining server replacement cycles is to base decisions on network performance data and forecasts of business needs.

Gartner Data Center Conference 2010 Coverage (Days 2-4)

As mentioned in the previous post, I wanted to compile the most interesting and valuable info from the keynote presentations (in the form of tweets, following the #GartnerDC hashtag conversation). A lot of the topics revolve around cloud computing, virtualization, and upcoming server hardware trends.

Highlights from Days 2-4

@nlyte: 65 percent used or plan to use outside help with DC consolidation #GartnerDC

@nlyte: 57 percent stated internal politics was the most difficult aspect of DC consolidation #GartnerDC

@ciscoDC: Gartner: 75% of server workloads virtualized by 2014 vs 25% today; 50% of all new workloads are virtualized #GartnerDC

@BarbaraRuane #APC perspective from #Gartner show – Power and cooling top of mind with decision makers #GartnerDC

@az990tony #GartnerDC [blog] Day 2 – Choosing the Right Storage for your Server Virtualization environment

@jayfry3 When choosing a #cloud provider, what matters most to you? @cloudpundit‘s #gartnerdc audience says “best-in-class security.”

@TonyKnowsPower: #GartnerDC Step away from the hardware – think in terms of Service Availabilty not server, storage & network uptime.

@SachaAtMsft MacDonald: Cloud computing is forcing us all to understand where information exists at any point during its lifetime. #GartnerDC

@shorinsean 39 % of conference attendees prohibit access to social media at work. #gartnerDC

@nlyte In 2011 which big 5 provider will you purchase tech from that you are not today? 33%Cisco 9%Dell 18%HP 17%IBM 8%Oracle 15%none #GartnerDC

@nlyte Understanding Data Center Transformation and the “Big Five” Providers: Cisco, Dell, HP, IBM and Oracle #GartnerDC


Gartner Data Center Conference 2010 Coverage

The Gartner Data Center Conference 2010 is underway in Las Vegas this week, December 6-9. This show has a lot of great information on the latest data center trends including new server hardware, blade enclosures, and enterprise servers. For those that could not attend the event, below is a compilation of the most interesting and valuable info from the keynote presentations (in the form of tweets, following the #GartnerDC hashtag conversation).

Day 1:

@FinancialPath: Heavy emphasis on Cloud Computing, Virtualization and Storage at #GartnerDC with Financial Services often being the first Vertical mentioned

@samcharrington 66% of attendees pursuing private cloud computing by 2014. #GartnerDC

@shorinsean Red Hat Linux by far the most popular Linux flavor in the data center for attendees. #gartnerdc

@nlyte #GartnerDC live poll of approx 2,000 attendees reported #1 issue faced by the data center today is management of power, cooling and space

@AppDynamics Donna Scott: Do a biz impact assessment on your apps — 20% of mission-critical apps should get 80% of investment.#GartnerDC

@adamfazio heard another new term I like: P2V2C – physical to virtual to cloud. #gartnerdc

@TonyKnowsPower: How times have changed blades are now referred to as traditional servers #gartnerdc

@shorinsean: Interesting facts from #gartnerdc keynote: 2.6B Google searches per month in 2007, 31B last month. That’s a 1200% increase.

@Isilon: Gartner says data growth over the next 5 years will be 800% & 80% of data is unstructured. #GartnerDC

@Datapac_ICT “The only bad cloud strategy is no cloud strategy” says @tombitt at #GartnerDC

@Datapac_ICT Gartner: 75% of server workloads virtualized by 2014 vs 25% today; 50% of all new workloads are virtualized #GartnerDC

Check back tomorrow for coverage of Day 2…

Securing Data Center Operations with Virtualization and Private Clouds


Secure Your Data Center Operations Before Virtualization

Organizations transitioning to a virtualized or cloud IT model need to invest in security strategy, technology, organization and skills. Vendors need to provide better integration between security and service management plus security tools to better support heterogeneous virtualized and physical environment.

Key points in regards to security virtualization include:

· The major driver of virtualization is the efficiency improvement of data center operations. The least important drivers are preparation for Cloud IT, which is closely followed by meeting Green IT targets.

· The major inhibitor for implementing virtualization security is the lack of expertise and skills to plan and implement it. Around a quarter of organizations believe that virtual environments in general are less secure than physical environments.

· The most important security challenge and concern is around “data sprawl”. This issue is closely followed by concerns relating to the fulfillment of both regulatory compliance and internal audit requirements in a virtualized or cloud environment.

· One key issue highlighted by the survey is that nearly three quarters of respondents are concerned about the far-reaching privileges introduced by hypervisors which might lead to abuse. There are technologies available today to mitigate the risks posed by privileged access in virtualized environments, but these are not widely deployed.

· There is a lack of integration between virtualization, security, and service management with less than half the organizations reporting any type of integration in this area.

· Too many security activities are still dependent upon manual processes, these processes are performed without supporting technology and the scale of virtualization makes this approach untenable.

· By not investing in virtualization security when there are well identified security threats, organizations are taking unnecessary risks which could easily be mitigated.

How does your organization plan to address risks of important data in virtualized environments from falling into unsecured environments?