By Kevin O’Toole
Published in the February 2012 issue of Today’s Facility Manager
From cloud computing and storage to virtualized applications and disaster recovery there are a variety of reasons companies are transitioning applications from “edge based” servers operating in a company’s local facility to centralized data centers. This transition to data centers means that facility managers (fms) are engaging with their Information Technology (IT) colleagues on new types of issues. The types of issues they must work together to solve are evolving from ensuring there is sufficient space and power to operate local servers, to ensuring there is excellent network connectivity to reach applications and storage located in remote data centers. As the cloud and virtualized services continue to expand in popularity, fms who are knowledgeable about the network needs of such services will be better business partners to their IT colleagues.
Five Critical Data Center Connectivity Questions
Making the move to a specialized data center can reduce capital and operational expenses and enable businesses to increase their application and storage capacity easily as their needs change. The benefits of using industrial strength data centers are abundant, but without reliable, flexible, and scalable high bandwidth connectivity, explosive network traffic can negatively impact application performance. As they work more closely with IT departments on these issues, fms are learning the impact that network connectivity can have on daily operations. Network connectivity is a critical component of a data center strategy and can play a key role in a company’s success. As fms work with their IT colleagues, they can evaluate network connectivity to the data center and determine if it can support their current and future needs.
Here are five important questions for fms to consider when selecting a data center and associated network provider:
1. Can the data center network accommodate my current and future bandwidth requirements?
Bandwidth hungry applications and services (such as cloud computing, business continuity/disaster recovery, business process automation, Software-as-a-Service [SaaS]), and rich media require high performance, reliable connectivity to work properly. Fms should make sure that network connections to the Internet and other facilities can provide the bandwidth levels they need, such as 100 Megabits per second (Mbps), 500 Mbps, or even 1 Gigabit per second.
It is also important to determine whether the network provider offers dedicated bandwidth and different classes of service. Offerings such as these deliver reliability and ensure companies’ have the network performance they need, when they need it. The ability to purchase bandwidth in flexible increments is also ideal, as it allows companies to customize solutions that meet the specific requirements of a given site or application.
Once it’s determined that a network provider can support a company’s current bandwidth needs, the next question must be this—what happens if bandwidth requirements change?
Fms and their IT counterparts should consider network providers that not only allow them to purchase the amount of network bandwidth they need, but also offer the option to scale up easily as the company grows. Preference should be given by fms to network operators capable of delivering additional bandwidth through remote, electronic provisioning as opposed to those whose networks may unfortunately require time consuming and costly on-site network reconfigurations.
2. How often does the network experience connectivity disruptions?
Network availability is a critical factor in data center connectivity—as applications and storage centralize, a network outage which severs connectivity to those applications has a greater impact on business operations. Repeated outages can have substantial impact on employee productivity and company reputation.
Facilities and IT leaders should, therefore, evaluate the percentage of time a carrier’s network is available, the average time it takes a network provider or carrier to respond to an outage, and the average time it takes to restore service. The selected carrier must have the financial strength to invest in both the network itself and the operational systems necessary for proper network monitoring and maintenance.
3. Are the network facilities truly diverse from other carriers?
It is common for a data center or other building to be served by both the Incumbent Telephone Company (e.g., Verizon) as well as a number of Competitive Local Exchange Carriers (CLECs). While this may provide the appearance of “last mile” network diversity, in reality, many carriers rely on reselling each other’s networks. What may appear to be a diverse network solution purchased from two different carriers may in fact simply be the same facilities traveling to same telephone company central office.
When selecting multiple carriers to build network redundancy, facility and IT leaders should ask hard questions about which company actually owns and operates the underlying network. Ensuring the network is truly diverse and owned/operated by the carrier selling the service delivers both true redundancy and allows the carrier to provide more timely service restoration in the event of an outage. Carriers that rely on third-party network facilities (e.g., by reselling telephone company networks) often must ask the telephone company or another underlying carrier to perform any actual network repair. Relying on a carrier that owns its network eliminates this middle man during outage restoration.
4. Does the data center network have the reach that is needed?
It doesn’t matter how scalable, flexible, and reliable a network may be if it doesn’t have the reach to accommodate a company’s needs. When determining a network’s reach, fms may find the following questions helpful:
- How many major markets does the network reach, and what are they?
- Is network infrastructure deployed
- in strategic locations for maximum coverage?
- Is the network wholly owned by one provider?
- Is the network physically separate from traditional telecommunications providers? If not, disruptions to the telco network can impact the connection to the data center.
5. What technology is being used to provide the network services?
The choice of the actual network technology used to provide connectivity is critical, as this will determine the network’s long run financial and operational characteristics. While there are many legacy choices available from incumbent carriers (e.g., SONET, T-1s lines, or ATM), more companies are finding that Metro Ethernet is the most efficient and cost-effective way to provide high performance connectivity to a data center. Metro Ethernet is significantly faster than T1, Frame Relay, and dedicated lines, and it has become the preferred solution for enterprises, according to a recent study conducted by analyst firm Vertical Systems Group.
Moving Data Centers Into The Future
Facility and IT leaders should look for Metro Ethernet services that have been certified by the Metro Ethernet Forum. Standardized, certified services ensure that it will be easier to interconnect Metro Ethernet services from multiple providers to help build a diverse network. Further, standards based solutions have a history of driving down costs over the long run (as competitors are forced to deliver superior service at lower prices rather than locking their customers’ into proprietary solutions which become difficult to evolve or switch to another provider).
For many fms, deciding to move IT operations to a dedicated data center is a big step. Selecting a network provider that delivers high performance, reliable connectivity to the data center is critical to unlocking the productivity and cost improvements of hosted, cloud-based services.
Fms who work together with their IT colleagues and ask the right questions about a network provider’s performance, flexibility, and reliability should be better able to select the best solution for their organizations. For many companies, the most appropriate answer to these questions may indeed be something completely new.
O’Toole serves as senior vice president of product management and strategy for Comcast Business Services. In this role, O’Toole is responsible for the introduction and life cycle management of data, telecom, video, and communications products serving businesses and mobile/cellular carriers.
Other posts by