Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Sign In with OpenID
Advertise on LowEndTalk.com

In this Discussion

Availability Zones, anyone?

Availability Zones, anyone?

concerto49concerto49 Member
edited October 2012 in General

Amazon EC2 has availability zones per region. Would anyone be interested in this from a LEB perspective?

Most providers have different regions - some have 2, some have lots - but just interested if there would an interest in availability zones and/or if there is, why isn't it being offered?

Internap VPS, Web Hosting and more - Cloud Shards | Need a VPS Upgrade?
Query Foundry, LLC AS62638

Comments

  • bdtechbdtech Member
    edited October 2012

    AWS "Availability Zone" = Separate Datacenters in close proximity. Which is at most a several miles apart (low latency connection, redundant power/network providers, etc..)

  • Same datacenters even, just other cages.

    Opinions/Posts are to be assumed my own/personal and not company related unless obvious
    Currently unemployed | Available for consulting | http://as198412.net | https://william.si

  • bdtechbdtech Member
    edited October 2012

    @William Not at AWS. They are completely separate.

    Q: How isolated are Availability Zones from one another? Each availability zone runs on its own physically distinct, independent infrastructure, and is engineered to be highly reliable. Common points of failures like generators and cooling equipment are not shared across Availability Zones. Additionally, they are physically separate, such that even extremely uncommon disasters such as fires, tornados or flooding would only affect a single Availability Zone. http://aws.amazon.com/ec2/faqs/#How_isolated_are_Availability_Zones_from_one_another

  • Then the extra redundancy seems marginal compared to separate regions. Construction cuts the big line, they're still all down. Regions at least provide real redundancy against disasters.

  • JarJar Member
    edited October 2012

    @averell said: Then the extra redundancy seems marginal compared to separate regions. Construction cuts the big line, they're still all down. Regions at least provide real redundancy against disasters.

    That would be my thought as well. My observation is that the LEB crowd mostly prefers their own redundancy solutions utilizing many small VPS systems, setting aside the obvious demand for local redundancy inside the node. The reason for this being that providers seem to have difficulty matching it at LEB prices, and even though the cost of several VPS does add up, they're still not at the mercy of a single provider.

  • @averell said: Then the extra redundancy seems marginal compared to separate regions. Construction cuts the big line, they're still all down. Regions at least provide real redundancy against disasters.

    The idea is, I think, e.g. if you had a web application cluster, database cluster or something that requires multiple applications to talk to each other with very low latency you will find it useful. You want redundancy but without the latency being high, otherwise e.g. they can be out of sync anyway when they replicate. For example, Amazon does this for their own Relational Database offering.

    Not sure what other uses there are.

    Internap VPS, Web Hosting and more - Cloud Shards | Need a VPS Upgrade?
    Query Foundry, LLC AS62638
  • I agree with @william here.

    Same Datacentre hooked up to different power supply lines and connections etc.

    physically separate is so vague it could be anything.

    Technically all of my VPS nodes are physically separate (different floors and suites) and have A/B power connections with two generators etc etc. therefore I could advertise as having availability zones under the vagueness of the FAQ.

    HTTP Zoom - httpzoom.com
  • @httpzoom lol "additionally, they are physically separate, such that even extremely uncommon disasters such as fires, tornados or flooding would only affect a single Availability Zone."

    Your zones are tornado redundant?

  • Yep, by being in a place where tornado's are extremely rare!

    HTTP Zoom - httpzoom.com
  • Fires too? I'm positive they use multiple DCs in "regions"

  • @bdtech said: Fires too? I'm positive they use multiple DCs in "regions"

    I was wondering if people would in interested in multiple DCs in "availability zones", e.g. 3 different DCs in New York or Texas etc. As in the same state, but different location.

    Internap VPS, Web Hosting and more - Cloud Shards | Need a VPS Upgrade?
    Query Foundry, LLC AS62638
  • @bdtech said: @httpzoom lol "additionally, they are physically separate, such that even extremely uncommon disasters such as fires, tornados or flooding would only affect a single Availability Zone."

    Your zones are tornado redundant?

    Didn't an entire Amazon AWS region go offline recently due to tornados?

    Appreciate my posts/software/guides? Donate (PayPal/Flattr/Bitcoin): http://cryto.net/~joepie91/donate.html | irc.freenode.net #lowendbox

  • @joepie91 said: Didn't an entire Amazon AWS region go offline recently due to tornados?

    Ireland lost power..... twice.

    Opinions/Posts are to be assumed my own/personal and not company related unless obvious
    Currently unemployed | Available for consulting | http://as198412.net | https://william.si

    Thanked by 1klikli
  • InfinityInfinity Retired Staff

    Ireland had a tornado?

    我是一个巨魔 (;

    Thanked by 1klikli
  • @Infinity said: Ireland had a tornado?

    Romania has few tornadoes (but they are more and more common as the warming is progressing) but still has plenty of power failures. I recently lost my 100 days uptime at work because of a power failure that lasted more than an hour. Sure, big DC have own generators and stuff, but still data cables can be cut by an earthquake, gas explosion, drunk digger even if they are deep in the ground, also underwater cables are even more at risk. There is no such thing as a sure thing, we may not even be sure if the sun will rise again tomorrow, but to take extreme measures to protect against such events when they only occur for a few minutes a year in average is kinda not worth it, the redundancy itself might introduce a new point of failure if you make a mistake, and humans make mistakes more often than an earthquake can cut some cables. M

    top - 22:32:38 up 906 days, 2:58, 1 user, load average: 3.94, 5.13, 8.38

Sign In or Register to comment.