Data Center

The Million Dollar Question – Part 1

 

You’ve heard it before in interviews, sales meetings, political debates, and even on The Bachelorette. You’ve even asked it before when the decision was yours to make. You tee your prospect up with that classic question that basically invites them to leave it all on the table. Every ounce of drive to make the connection pours into this well-crafted, often thought of, highly rehearsed final petition.

When someone hears it, they know the bases are loaded. You have one last chance to knock it out of the park, or no one is going home happy.

You toss the pitch right to their sweet spot. It’s make it or break it, and then it’s out. “Why should I choose you?

Please allow me to swing away and tell you how we answer that question. When comparing data center providers you can expect that most will all have some things in common.

Power? Check

We’ve got the power! After all, why on earth did you choose to send your servers away to boarding school at FIBERTOWN prep? Because after the umpteenth time your building lost power to a forlorn and heartbroken squirrel who sacrificed himself to the local substation, you finally decided to make keeping the lights on someone else’s problem. FIBERTOWN standardizes on dual path power with independent and concurrently maintainable power distribution elements all the way through from the utility provider to the cabinet. This means one side can be maintained while the other side keeps you running. We plan to be up and running long after the zombie apocalypse has claimed everyone except the small contingent of American crossbow aficionados.

Cooling? Check

Don’t forget to pack your hoodie as you conveniently decide to complete that annual data center hardware audit you were supposed to do back in February.

Visiting a data center during July in Texas is like taking a dip in the pool, but you can justify doing it on the clock. Sufficient cooling in your tier II facilities is nice, but your more fault tolerant upper tier data centers will have room to compensate for a device failure.

FIBERTOWN employs fault tolerant installations of chilled water systems, CRAH/CRAC units (it means computer room air conditioner, we do not deal in narcotics), and resilient chilled water distribution systems all geared to keeping you and your critical infrastructure nice and chilly.

Security? Check

Levels of security will vary from perimeter fences and obscure, unmarked buildings to dual factor authentication with biometric scans and 24x7x365 on-site security staff. Sometimes you could see gun turrets, high fences with razor wire and armed personnel should you wish your servers, routers, and switches are securely incarcerated. These are easy line items to check off.

Serious facilities will lean towards the more secure methods including full-time security staff, solid locked down facilities, dual factor authentication and extensive monitoring of entry points to maintain accountability. The FIBERTOWN server zoo has no intentions of letting the lion out of its cage.

Connectivity? Check

Even tier II designed data centers will have bandwidth in spades. With your favorite carriers riding other carriers the last mile, you can bet you get what you need for connectivity. Both FIBERTOWN data centers sit on a nexus of fiber connectivity, otherwise we may as well have named ourselves COPPERTOWN.

So what is the difference? Are all data centers alike? You are deliberately avoiding answering the million dollar question!

There is one more key differentiator. Can you guess what it is? Stay tuned and check out part two of The Million Dollar Question!

Data Center

How to Build a Highly Available Network

 

There are multiple ways to build a highly available network. Let’s assume you require a 100% concurrently maintainable network, which ensures when any one connection is lost you can maintain network connectivity.

Internet Service Provider

If you maintain multiple connections that end users must access 24×7, you can achieve connectivity through DNS and BGP.

Adjust Domain Name Servers (DNS)

Consider what happens if you have two separate carriers providing diverse IP subnets and your end users are accessing them via DNS and your primary carrier fails? Your users will attempt to access unavailable services.

To avoid those situations, set the time-to-live on DNS entries to five minutes ahead of time, which will allow you to change the IP addressing. As the updated DNS propagates, it will allow your users to slowly adjust.

There are automated DNS services such as UltraDNS, which monitor external services, detect a failure and automatically adjust DNS entries. The advantage of an automated system is that it performs failover at any time and you can expect a failover of roughly 5 minutes.

 

For more information on network connectivity
<< Download our free Data Center Design Guide >>

 

Border Gateway Protocol (BGP)

BGP is a dynamic routing protocol that controls the flow of information on the Internet. BGP allows you to influence the flow of traffic both in and out of your network and is preferable to DNS management.

An Autonomous System Number (ASN) is assigned via a local registry. This ASN allows you to peer with upstream providers. You can advertise your registry-assigned IP addresses with your primary and secondary ISPs. If one ISPfails, the identical addresses are available via your secondary ISP.

Hardware

Multiple connections should be hosted on multiple routers to maintain connectivity should a single router fail. Border routers can then be connected to redundant firewalls. From the firewall, you should have connectivity into a pair of core switches for layer 2 aggregation.

Can you see the pattern? For max redundancy, host connectivity on multiple devices from beginning to end. Often this configuration can be collapsed where multiple functions are performed via a single piece of equipment. Your required uptime should be balanced with cost.

Data Center

5 Things about Colocation Your Boss Wants to Know

When companies consider colocation as an option to upgrade space and power while eliminating downtime, they start evaluating potential data center providers. When reaching out to get a quote or visit a facility, there are 5 things you need to know in order to answer the 5 questions your boss will ask. Here’s what they will want to know and five ways to make sure to get answers.

1. How much is this going to cost?

2. How safe is our data?

3. Will my bandwidth be slower/faster/go down?

4. Who has access to our equipment?

5. What legal ramifications are we responsible for?

 

For more questions to ask when evaluating data centers << Download our free Colocation Checklist >>

 

1 Have a clear understanding of your budget and downtime requirements

Consider your budget and the “hidden costs” included in colocation contracts. These include cage setup fees, circuit installation fees, bandwidth, cross connects and IP addresses. Make sure you have a clear understanding of what you are paying for and why.

Downtime is another area to consider. How much can your business afford to be down? If an outage is only a mild headache, then perhaps a Tier II data center would suffice. If downtime costs you money, a Tier IV high availability data center is worth the premium charge.

When discussing service level agreements (SLA) and uptime, always ask how many times they have gone down and had to pay a customer. Just because a data center advertises a 100% uptime SLA doesn’t mean they never had to pay customers due to downtime.

2 Determine acceptable levels of security

Security is always a top issue when evaluating a data center. Physical security such as site location, perimeter monitoring, facility access, vendor management and video recording backup are important to assess.

SANS Institute issued a checklist www.sans.org that is helpful to review. Ask about the SSAE 16 (SAS70) certification to make sure compliances are being met.

Knowing what compliances are needed such as HIPPA and PCI help ensure the data centers willingness to allow auditors the ability to verify proper regulations and controls are in place.

3 Determine connectivity requirements and carrier service

The key to connectivity is having a carrier-neutral data center. If you are currently working with a telecom carrier, make sure they’re connected to that data center.

Facilities with many telecom providers offer lower Internet costs, MPLS networks, point-to–points and PRIs for PSTN termination and carrier cross connects.

Consider what options the data center can offer you. Many offer a blended product from two separate providers, rather than from a single carrier, which would be considered a single point of failure.

4 Inspect the data center support team

Knowing the type of support offered is important. This can save you time and money when it comes to sending one of your staff to do simple tasks such as rebooting services or tape rotation.

Do they offer 24x7x365 support? Is that support located onsite or through a 3rd party? Having local onsite support is key to a quick response time on any hands-on needs you may have. Some data centers offer remote hands while others offer only reboots.

Determine all service levels you require, skill level and certifications of staff, and the cost per hour for utilization.  This is critical if the site is not geographically near you.

5 Managed service agreement and service level agreement negotiations

Make sure to carefully review the MSA and SLA with your company attorney. Verbal negotiations and agreements will not hold much standing as data centers strictly rely on contract terms in the agreement.

Make sure to review all sections and pricing structures to avoid any issues or misunderstandings.

 

For more questions to ask when evaluating data centers << Download our free Colocation Checklist >>