5 ms·

Why Amazon Reserved Instances don’t make economic sense for startups

34 points by michaelkscott 14y ago

ubercore 14y ago
One key part (mentioned by a comment in the story itself) that wasn't clear to me at first is that you're not purchasing a single reserved instance. The reservation applies to any instance of the type you've reserved. So regardless of specific instances running 24/7, it can still make sense if you generally have a specific instance flavor running all the time. That makes it a lot more flexible, and a better value in a lot of cases, I think.
- nirvdrum 14y ago
  Just keep in mind that the reservation is also for a particular availability zone. You need to stay within that zone to get the discounted rate.
  - ubercore 14y ago
    Was just checking my comments, and realized I recognized your username. WPI?
    - nirvdrum 14y ago
      Indeed. Sorry, don't recognize your username and your profile is a bit thin. I graduated in '04 and '06.
      - ubercore 14y ago
        Lived in Stoddard (B I think?) freshman year.
moe 14y ago
I'll extend and restate the obvious: EC2 doesn't make economic sense for most startups. EC2 shines for very small, very large or very dynamic deployments (jitter of at least +/- 20 servers). Most startups outgrow "very small" (2-4 Servers) quickly and then stick around in the mid-range (5-30 servers) for an eternity. During this mid-range phase EC2 is usually 2x-3x more expensive than rented dedicated servers.
- ceejayoz 14y ago
  > During this mid-range phase EC2 is usually 2x-3x more expensive than rented dedicated servers. That expense might be worth it, though, if your startup goes unexpectedly viral and the site goes down for hours or days while you wait for your host to provision new dedicated servers for you.
  - mceachen 14y ago
    But in this sort of timeframe, you'd be able to switch your DNS to point to EC2 or heroku or whatever and scale out someplace else.
    - sandieman 14y ago
      What would you suggest instead?
- jasonwatkinspdx 14y ago
  1.) If you want to use cloud providers for your infrastructure and you get big, you need to get to where you're multi-homed and can leverage competition in discounts. You think netflix pays those rates? 2.) For small systems, don't forget hybrids! EC2 is a _steal_ as a disaster contingency option. Run your own stuff. Replicate state to EC2. Keep your configuration managed and up to date both places. Build your routing horizon (dns most likely, perhaps anycast) to be able to shift traffic between the two fluidly and quickly. Moron at your data center kills the power? No big deal, spin up instances and keep users happy. Got a big but occasional data processing job? No problem, and less ways the job can affect the production system. The open source tooling for doing this kinda sucks but it's getting better fast. It's a huge opportunity for smaller businesses to level advantage vs large organizations with multiple in house data centers.
- rickmb 14y ago
  Which would be relevant if the cost of bare servers wasn't just a minor part of the cost of maintaining the infrastructure for most organizations.
crb 14y ago
This article is from 2009, and while many of the points it makes are still valid, there is at least one that isn't: > As a result, it ONLY makes sense to consider a dedicated instance for a machine that will be running 24 hours a day, 7 days a week. Amazon have addressed this with their Light, Medium and Heavy RIs. The break-even points on the cost savings are on this page: http://aws.amazon.com/ec2/reserved-instances/ http://aws.amazon.com/ec2/reserved-instances/ The one thing I would add about RIs is: they only apply to a pairing of AZ and instance size. They suit people who scale out (adding more instances) and don't suit people who scale up (changing their instance sizes and not using the old reservations).
Isofarro 14y ago
Site seems to be down. So Google cached version: http://webcache.googleusercontent.com/search?q=cache:67WG0Qe1XLIJ:www.jonathanboutelle.com/why-amazon-reserved-instances-dont-make-economic-sense-for-startups&strip=1 http://webcache.googleusercontent.com/search?q=cache:67WG0Qe...
Isofarro 14y ago
I went along to Amazon AWS Cloud for startups and developers (http://aws.amazon.com/about-aws/AWS-cloud-for-startups-developers-London/ http://aws.amazon.com/about-aws/AWS-cloud-for-startups-devel...) last week in London. One very interesting talk was http://www.slideshare.net/AmazonWebServices/optimizing-your-infrastructure-costs-on-aws http://www.slideshare.net/AmazonWebServices/optimizing-your-... -- same slide deck, but different speaker in the London event. If there's a video available of this talk, try to watch it. On Reserved instances a one year Reserved instance is equivalent to 6 months running a non-reserved instance. So it starts to save you money after 6 months. Don't remember if this was a High, Medium or Low usage reserved instance. (slides 15, 16, 17) Other cost savings techniques I found interesting: * Using AWS ancillary services instead of building your own on EC2 servers saves a big chunk of cost (elastic load balancer, SQS, DynamoDB) -- slides 30-34 * Consider bidding for spot instances if you have data processing requirements that fit those usecases (Map/Reduce jobs for example). This will save you heaps. -- slides 21-30 * It's really all about fitting your usage graph to the right combination of EC2 instances, so something long-term to handle the average traffic. A medium resource to handle the daily peak loads, and something extra on hand to handle a short but sharp spike in traffic. And spot instances to push down the per hour resource cost in batch processing. (Disclaimer, I work for LOVEFiLM, an Amazon company, but not involved in AWS side of things)
- sp332 14y ago
  It tells you right on this page exactly how much of a year you'd need to run each instance type to break even on the reservation. https://aws.amazon.com/ec2/reserved-instances/ https://aws.amazon.com/ec2/reserved-instances/
sauravc 14y ago
Not only is this article is outdated, it makes broad sweeping generalizations that make little sense for usage patterns that differ from the author's. Reserved instances make perfect sense if your server on EC2 is running 24/7 for months on end. In fact, with now that Amazon has different rates for medium/light/heavy utilization, it makes perfect sense for even more situations. We've taken advantage of RI's at my startup and the benefits are pretty clear. Our app servers are micro instances running PHP, Python and Java (SOLR). We scale by adding more instances not by renting bigger servers, so our RI's are always in use. It's also worth noting that the reserved instances for RDS are a great way to reduce costs as well (if you choose to use RDS).

34 points by michaelkscott 14y ago

One key part (mentioned by a comment in the story itself) that wasn't clear to me at first is that you're not purchasing a single reserved instance. The reservation applies to any instance of the type you've reserved. So regardless of specific instances running 24/7, it can still make sense if you generally have a specific instance flavor running all the time. That makes it a lot more flexible, and a better value in a lot of cases, I think.

nirvdrum 14y ago
Just keep in mind that the reservation is also for a particular availability zone. You need to stay within that zone to get the discounted rate.
- ubercore 14y ago
  Was just checking my comments, and realized I recognized your username. WPI?
  - nirvdrum 14y ago
    Indeed. Sorry, don't recognize your username and your profile is a bit thin. I graduated in '04 and '06.
    - ubercore 14y ago
      Lived in Stoddard (B I think?) freshman year.

moe 14y ago

I'll extend and restate the obvious: EC2 doesn't make economic sense for most startups. EC2 shines for very small, very large or very dynamic deployments (jitter of at least +/- 20 servers). Most startups outgrow "very small" (2-4 Servers) quickly and then stick around in the mid-range (5-30 servers) for an eternity. During this mid-range phase EC2 is usually 2x-3x more expensive than rented dedicated servers.

ceejayoz 14y ago
> During this mid-range phase EC2 is usually 2x-3x more expensive than rented dedicated servers. That expense might be worth it, though, if your startup goes unexpectedly viral and the site goes down for hours or days while you wait for your host to provision new dedicated servers for you.
- mceachen 14y ago
  But in this sort of timeframe, you'd be able to switch your DNS to point to EC2 or heroku or whatever and scale out someplace else.
  - sandieman 14y ago
    What would you suggest instead?
jasonwatkinspdx 14y ago
1.) If you want to use cloud providers for your infrastructure and you get big, you need to get to where you're multi-homed and can leverage competition in discounts. You think netflix pays those rates? 2.) For small systems, don't forget hybrids! EC2 is a _steal_ as a disaster contingency option. Run your own stuff. Replicate state to EC2. Keep your configuration managed and up to date both places. Build your routing horizon (dns most likely, perhaps anycast) to be able to shift traffic between the two fluidly and quickly. Moron at your data center kills the power? No big deal, spin up instances and keep users happy. Got a big but occasional data processing job? No problem, and less ways the job can affect the production system. The open source tooling for doing this kinda sucks but it's getting better fast. It's a huge opportunity for smaller businesses to level advantage vs large organizations with multiple in house data centers.
rickmb 14y ago
Which would be relevant if the cost of bare servers wasn't just a minor part of the cost of maintaining the infrastructure for most organizations.

crb 14y ago

This article is from 2009, and while many of the points it makes are still valid, there is at least one that isn't: > As a result, it ONLY makes sense to consider a dedicated instance for a machine that will be running 24 hours a day, 7 days a week. Amazon have addressed this with their Light, Medium and Heavy RIs. The break-even points on the cost savings are on this page: http://aws.amazon.com/ec2/reserved-instances/ http://aws.amazon.com/ec2/reserved-instances/ The one thing I would add about RIs is: they only apply to a pairing of AZ and instance size. They suit people who scale out (adding more instances) and don't suit people who scale up (changing their instance sizes and not using the old reservations).

Isofarro 14y ago

Site seems to be down. So Google cached version: http://webcache.googleusercontent.com/search?q=cache:67WG0Qe1XLIJ:www.jonathanboutelle.com/why-amazon-reserved-instances-dont-make-economic-sense-for-startups&strip=1 http://webcache.googleusercontent.com/search?q=cache:67WG0Qe...

Isofarro 14y ago

I went along to Amazon AWS Cloud for startups and developers (http://aws.amazon.com/about-aws/AWS-cloud-for-startups-developers-London/ http://aws.amazon.com/about-aws/AWS-cloud-for-startups-devel...) last week in London. One very interesting talk was http://www.slideshare.net/AmazonWebServices/optimizing-your-infrastructure-costs-on-aws http://www.slideshare.net/AmazonWebServices/optimizing-your-... -- same slide deck, but different speaker in the London event. If there's a video available of this talk, try to watch it. On Reserved instances a one year Reserved instance is equivalent to 6 months running a non-reserved instance. So it starts to save you money after 6 months. Don't remember if this was a High, Medium or Low usage reserved instance. (slides 15, 16, 17) Other cost savings techniques I found interesting: * Using AWS ancillary services instead of building your own on EC2 servers saves a big chunk of cost (elastic load balancer, SQS, DynamoDB) -- slides 30-34 * Consider bidding for spot instances if you have data processing requirements that fit those usecases (Map/Reduce jobs for example). This will save you heaps. -- slides 21-30 * It's really all about fitting your usage graph to the right combination of EC2 instances, so something long-term to handle the average traffic. A medium resource to handle the daily peak loads, and something extra on hand to handle a short but sharp spike in traffic. And spot instances to push down the per hour resource cost in batch processing. (Disclaimer, I work for LOVEFiLM, an Amazon company, but not involved in AWS side of things)

sp332 14y ago
It tells you right on this page exactly how much of a year you'd need to run each instance type to break even on the reservation. https://aws.amazon.com/ec2/reserved-instances/ https://aws.amazon.com/ec2/reserved-instances/

sauravc 14y ago

Not only is this article is outdated, it makes broad sweeping generalizations that make little sense for usage patterns that differ from the author's. Reserved instances make perfect sense if your server on EC2 is running 24/7 for months on end. In fact, with now that Amazon has different rates for medium/light/heavy utilization, it makes perfect sense for even more situations. We've taken advantage of RI's at my startup and the benefits are pretty clear. Our app servers are micro instances running PHP, Python and Java (SOLR). We scale by adding more instances not by renting bigger servers, so our RI's are always in use. It's also worth noting that the reserved instances for RDS are a great way to reduce costs as well (if you choose to use RDS).