University Data Center
UDC: Frequently Asked Questions
Question: How much of the university’s overall server population can fit in the new UDC?
Answer: That’s difficult to say, but as current customers evaluate what they want to move, they find they have many options that can reduce the server and data storage footprint on campus: virtualization, more efficient physical servers, and converting to central services like Austin Disk and Austin Exchange Messaging Service. Often, departments can also retire aging switches and other old network equipment where the data center provides that functionality in top-of-rack switches.
Question: What is the life expectancy of the new UDC?
Answer: In general, the physical life expectancy of a data center building is different from its useful life expectancy as a data center. The latter will be dependent on the university’s computer utilization, growth and power requirements for newer server technology. With the current trajectory, we would project that the data center would not reach capacity for approximately 3-5 years.
The building is designed for redundancy to allow proper maintenance to take place, which will extend its useful life. In some facilities (including our current data centers), you operate in a “run-until-fail” mode. You can’t take equipment offline, so you plan for its inevitable failure. Our new building isn’t like that. It is designed to allow concurrent maintenance without affecting uptime.
Question: What kind of equipment can go into the new UDC?
Answer: The West Hall and East Hall are designed to accommodate 19” rack-mountable devices. There is only a limited amount of space for non-rack-mountable equipment. There are also some hardware configuration issues that can affect whether equipment can achieve the full service levels. Please see the System Standards and the SLA for a complete description.
Question: How often will you do maintenance on the power and network?
Answer:Typically network maintenance is performed at least twice a year and can be more frequent depending on network needs. Power maintenance is performed routinely and is dictated by each component’s specific requirements. All planned maintenance will be scheduled ahead of time and customers will be given advance notice through approved communication channels such as email distribution lists and the ITS Planned Maintenance Calendar.
Question: How do I request access to the facility to get to my servers?
Answer: To request building / Data Center access to the facility, submit a request to email@example.com.
To support easier server management, there is a dedicated remote management network that will allow administrators to remotely access their servers as needed. The UDC operators are also available 24x7x365 to perform tasks like power cycling. Customers can call the UDC operations staff at 512-471-0007.
Question: What if a server goes down a vendor must provide onsite emergency maintenance?
Answer: Customers should submit a ticket to firstname.lastname@example.org and call the UDC operations staff at 512-471-0007 and inform them that the vendor is expected. Operators are available on a 24x7x365 basis.
Question: What support is available to prepare servers for the redundant network?
Answer: Network connectivity is provided from A and B sources to every server for redundancy. Each server equipped with dual network interface cards (NICs) is connected redundantly to both the A and B sides of the network via the top of rack switches. Servers can do down during scheduled maintenance and unplanned events if servers are redundantly connected but not properly configured to use the network equipment using link aggregation control protocol (LACP) bonding.
Question: What is the cost to move my servers to UDC?
Answer: The rate is calculated by multiplying the per-watt rate of $0.90 times the "nameplate" power capacity of a device, as listed in the Rackwise database used by ITS. The nameplate value represents the device’s maximum power capacity, we de-rate that value by 25% based on the assumption that most devices do not operate at maximum capacity
Annual Fee: [Device Power Capacity (Watts) x .75] x $0.90
Question: My grant already includes an overhead amount that is paid to the university. Shouldn’t the co-location fee be covered by that?
Answer: Because the dynamics for individual award agreements vary, and the degree to which individual deans distribute returned overhead to principle investigators in their units varies, for all questions about how your grant- specific costs relate to co-location services and fees, please contact Mike Cunningham, Director University Data Centers.