Categories
Data Center

Data Center Tiers

Tier 1 = Non-redundant capacity components. things such as power have single feeds.
99.671% Uptime
no redundancy
28.8 Hours of downtime per year.

Tier 2 = Tier 1 + Redundant capacity components.
99.749% Uptime
Partial redundancy in power and cooling
Experience 22 hours of downtime per year

Tier 3 = Dual-powered equipment and feeds for power, cooling, and essential services.
99.982% uptime (Tier 3 uptime)
No more than 1.6 hours of downtime per year
N+1 fault tolerant providing at least 72-hour power outage protection

Tier 4 = All components are fully fault-tolerant including uplinks, storage, chillers, HVAC systems, servers etc. Everything is dual-powered.
9.995% uptime per year (Tier 4 uptime)
2N+1 fully redundant infrastructure (the main difference between tier 3 and tier 4 data centers)
96-hour power outage protection
26.3 minutes of annual downtime.

If you want to read up on the N redundancies. https://en.wikipedia.org/wiki/N%2B1_redundancy

Categories
WISP

Thoughts on carrier redundnacy

Recently there have been discussions on some lists about carrier redundancy.  I figured I would sum up some thoughts and add my own,

In today’s world of consolidation, takeovers, and cost saving measures carrier redundancy is something one should pursue with due diligence.  Below are some questions to know about your existing provider and any future providers. If you know this you can compare the differences between two providers.  By knowing the answer to these questions you can add two carriers which will complement each other.

1.Where does my circuit go when it leaves my equipment? Look at this from a regional perspective.  Where does it travel in the city? Where does it travel to the next city?

2.Does the provider’s lines share conduit with other providers?  They might not know this, but if you have two providers you can compare routes.  If they are in the same conduit or in separate conduit in the same trench that might not be ideal.  A backhoe could take both out. Do they share space inside the path with other carriers as well? If so, this could cause issues with contract disputes, not paying bills, and other business-related functions.  Imagine if carrier A is sharing conduit with Carrier B. Carrier A goes out of business and holds the conduit contract.  Where does that leave Carrie rB?

3.Where is the entry point to the facility for the provider’s circuits? If both come into the same part of the building this could be a potential weak point.  Ideally one would enter from the north (or south or whatever) and the other would enter from a different direction.  Also, they would travel up different conduits on different sides of the building.  This way if something like a car crashes through the building may be one of them will be protected.

4.Does the provider farm any of your circuit out to a 3rd party?  This is good to know when problems arise and the finger pointing begins.

 

5.Use tools such as a “Looking Glass” to see if there are differences in routes.  If you have two backbone providers and they have very similar routes to reach the major sites (ie. Google, Yahoo, etc.) then you could open yourself up for problems with latency and packet loss should those paths become congested of fail.  Ideally, you want ProviderA to have different routes than ProviderB.  This way if something outside their network is causing issues it won’t have as big of an impact on your network.   Think of this as a road.  You might have two roads leaving your town, but you don’t want both of those roads taking the same path to get to the outside world.

 

Also look at this from your own equipment perspective. If you terminate all your circuits on a single router you are dependent on that router. Same goes for anything.  If everything comes in over the same ladder racks that are a point of failure.   If all your equipment is in the same room that is a point of failure.

Redundancy can be as diverse as you want to. It boils down to mitigating the risk.  If you know all the risks you can say “Yeah I am willing to bring my cross connects over a single ladder rack because the likely-hood of that rack failing is a risk I will take.”

 

Visit the MTIN Family of Blogs
www.mtin.net
www.indycolo.net
www.j2sw.com