Blogger social buttons

Monday, 24 June 2013

How do cloud price drops affect the bottom line? TripAdvisor and Pinterest

Effects of cloud price drops on TripAdvisor and Pinterest

Over the last year and a half, we counted 29 price reductions in cloud services provided by AWS, Google Compute Engine, Windows Azure, and Rackspace Cloud. Price reductions have a direct effect on cloud users, but given the usual tiny reductions, how significant is that effect on the bottom line?

Last year I wrote about cloud cost forecasts for TripAdvisor and Pinterest. TripAdvisor was experimenting with AWS and attempted to process 700K HTTP requests per minute on a replica of its live site, and Pinterest was growing massively on AWS. In the wake of the cloud providers’ price reductions, I revisited TripAdvisor and Pinterest to compare the old and new cloud cost forecasts.

In my first blog post I asked:
  1. How much would it cost to deploy System X on Cloud Y? For example, how much would it cost to host TripAdvisor on the AWS US-East cloud with On-Demand Instances? What if it used one-year Reserved Instances?
  2. What happens to costs when the system grows? For example, Pinterest has around 410TB of data on S3 — what if that keeps growing at a rate of 25 percent every month, as it has done in the last 10 months?

PlanForCloud, a free cloud cost forecasting engine, can help you concentrate on the cost components of your cloud strategy. I created a couple of deployments in PlanForCloud to explore these questions. The results show how important it is update your cost forecasts after price reductions.

Trip Advisor

1. TripAdvisor's "700K requests/minute" deployment on the AWS US-East cloud - how much would it cost?

I first looked at TripAdvisor to try to determine how much its 700K requests/minute on the AWS US-East cloud would cost. I took the following deployment specs from TripAdvisor's blog post detailing its experimental system:

- 270 x Hi-Memory XLarge (m2.xlarge) running 24 hours/day as front-end servers
- 70 x Hi-Memory XLarge (m2.xlarge) running 24 hours/day as back-end servers
- 32 x Hi-Memory Xlarge (m2.xlarge) running 24 hours/day as memcache servers

- 5  x Cluster Compute 8XLarge (cc2.8xlarge) running 24 hours/day as database servers, each having a 1TB EBS volume with 100 IOPS attached

- 5TB of EBS with snapshots to S3 every month for backups of all databases

Data Transfer:
- 145TB of data going out to users every month (around 200GB/hour)

TripAdvisor's deployment on AWS US-East modelled in PlanForCloud

In October 2012, if TripAdvisor used On-Demand instances for the above deployment, they would have expected to pay around $1.7M per year. Re-running the simulation again today, this has been reduced to $1.6M per year due to the price reductions.

If TripAdvisor used one-year Heavy Utilized Reserved Instances, in October 2012 they would have expected to pay around $0.9M per year (including upfront reservation fees). This figure currently stands at $0.7M per year due to the price reductions. The following table shows the percentage reductions for the two scenarios.

One year cost forecast of TripAdvisor

It's important to note that any AWS On-Demand price reductions are passed directly to users, while the upfront cost of Reserved Instances is due at the time of purchase, and it's rare for AWS to issue a partial refund when it reduces this upfront price if it’s past 90 days.


When I wrote my original blog, Pinterest had around 410TB of user data on AWS S3, and historic trends showed that this was growing by around 25 percent every month. In October 2012, Pinterest would have expected to pay around $470K per month after 12 months of growth if they used AWS S3 Standard Storage. Today this figure stands at $362K, thanks to the price reductions. The same amount of user data would have cost around $319K per month if Pinterest used S3 Reduced Redundancy Storage; this figure currently stands at $284K. The following table shows the percentage price reductions. It's interesting to see that S3 Standard Storage has been reduced more than Reduced Redundancy Storage.
Monthly Cost Forecast

Lessons Learned

From these two case studies, it's clear that the significance of cloud price reductions on a business’s bottom line depends on specific scenario and usage levels. PlanForCloud includes more than 12,000 prices across six cloud providers, and these prices change regularly. When you sign-up for PlanForCloud, we'll notify you automatically each time a price reduction affects your cost forecasts.

-- Ali
Technical Lead, PlanForCloud

Monday, 17 June 2013

Rackspace Australia cloud price cheat sheet

Rackspace today released its OpenStack-powered cloud platform in Australia. As close partners with Rackspace,  PlanForCloud, the free cloud cost forecasting tool, enables Australian users to simulate their future costs by using the Australian cloud prices.

Here are the prices for the Rackspace Australian cloud:

Storage - Cloud Files

Outgoing bandwidth without CDN: $0.20 per GB

Storage - Cloud Block


All prices exclude the Australian Goods and Services Tax.

Some of the press have picked up on the fact that Rackspace Australia is priced higher than the company’s US offering, but this is not uncommon. Other cloud providers also charge different rates in different countries, some even in different currencies.

Our database holds more than 12,000 price points from the six major cloud providers PlanForCloud supports - AWS, Rackspace, Google, Azure, HP, and SoftLayer. Imagine trying to price different cloud strategies and deployments on a spreadsheet! Log into PlanForCloud as a guest now.

PlanForCloud Product Manager

Tuesday, 4 June 2013

New Features - Support & Other Cost Modeling and Import from RightScale

If your enterprise is moving towards cloud adoption, you need to plan a cloud adoption strategy that includes proof of concept (POC) projects, risk and benefit analysis, performance benchmarks and cost forecasting for different deployment options.

PlanForCloud, a free cloud cost forecasting engine, can help you concentrate on the cost components of your cloud strategy. PlanForCloud enables you to design deployments and factor in servers, storage units, databases, data transfers, and even seasonal growth patterns. You can then run your numbers through our simulation engine, which factors in the latest cloud provider prices to create a detailed three-year cost report.

Earlier this year, we announced a new feature that enables users to import a snapshot of their AWS usage into PlanForCloud and forecast costs going foreword as they increase their usage. I am excited to announce three more new features today:

1 - Model Cloud Support Costs
Each IaaS cloud provider offers different support levels and charges based on different criteria. Some charge based on total cloud spend, some on per-instance usage, and some are completely inclusive. When you design your deployment with PlanForCloud, you can now model different support options and do what-if analysis:
PlanForCloud - Cloud Provider Support Plans

2 - Model Other Costs
A deployment is made up of many services and cloud resources, but you must also take into account other costs, such as staff, licenses, SaaS, and PaaS, in order to accurately compare deployment costs. With PlanForCloud, you can now consider any monthly costs and add or remove them from your deployment analysis. You can then assess how significant your other costs are in relation to your infrastructure and support costs:
PlanForCloud - Cloud Cost Reporting

For instance, in this analysis, we planned for seasonality peaks for the busy shopping season over the holidays, which accounts for the peaks in November and December. The May peaks are due to the AWS Reserved Instance yearly renewal costs, which this deployment is using. We’ve also planned for permanent growth in storage and data transfers due to increase in demand.

3 - Import from RightScale

With more than 65,000 customers, including the likes of Samsung, Sony, BBC, Zynga, Animoto, and PBS, RightScale is by far the leader in cloud management. Sitting at a level above public and private clouds, RightScale enables enterprises to manage resources through a single pane of glass. No matter which cloud providers you use - AWS, Google, Rackspace, Windows Azure, SoftLayer, and others - you can now import a snapshot of your resources from the RightScale dashboard into PlanForCloud and forecast future costs.
PlanForCloud - Import from AWS or RightScale

PlanForCloud is free, so sign up and start using it to forecast future costs and growth patterns and do what-if analysis.

Product Manager at PlanForCloud

Tuesday, 16 April 2013

Windows Azure goes GA with big pricing announcement

Azure matches AWS pricing: Is it really the same?

Excited yet? We are! Today, our partners at Microsoft announced that Windows Azure is GA as of today. That means that customers can now combine the Windows Azure PaaS offering with the IaaS offering to cover a wide range of use cases.

What makes this announcement even more exciting from our perspective is that Microsoft is now committing that Azure will “match Amazon Web Services prices for commodity services such as compute, storage and bandwidth”. This commitment applies not just now, but going forward. This opens the game up for those price sensitive cloud users who until now have been led to believe that AWS is always the cheapest option. Azure starts this with a reduction of 21% to 33% for its on-demand compute instances.

There are a couple of questions which we need to look at and answer:

1 - Apples to Apples?
How do the services offered by both cloud providers match up? Do the servers have the same CPU, Memory and local storage?

I have mapped out the different Cloud Services compute from Azure below and have matched them up to the equivalent AWS EC2 resources below.

AWS vs Azure - different instance types
As we can see from the above data table, the normal instance types offered by both cloud providers do match up quite well across compute units and memory. AWS offers slightly more memory (10-20%) in all instances except Micro. However, Azure offers more storage by default.

When it comes to High Memory instances, if we compare the Azure A6 and A7 instance types to AWS’ Hi-Mem 2XL and 4XL, we can see that AWS offers both more compute units and memory. However, AWS offer additional instance types. Perhaps Azure will look at adding these instance types in the future. If we compare the Azure A6 and A7 instance types to AWS’ Hi-Mem 2XL and 4XL, we can see that AWS offers both more compute units and memory. 

We need to note that, although these numbers tell one side of the story, the other side comes from running real workloads to assess performance. This matters because your application may run perfectly on a small Azure instance, but may struggle on a small AWS instance, or vice versa.

2 - AWS vs Azure costs - do they really match?

Azure has announced that they will match AWS prices. We need to look at the public price sheets to see if this is as a given or if this is something you would need to request from Azure after you have got your bill and usage.

Lets look at the public price tables:

AWS vs Azure - exact same on-demand prices
The cheapest prices (On-Demand in cheapest region, running Linux) all match 100% between Azure and AWS. Azure has now thrown down the gauntlet, stating they want to be as cheap as AWS so that users do not need to ask the question - who is cheaper. This is a really interesting approach to take. If Azure was to reduce their compute prices lower than AWS, I wonder if AWS would follow suit?

Update: It is interesting to note that AWS is marginally more expensive for Windows instances by the third decimal point. For example, a large instance running Windows on Azure costs $0.36 per hour, and will cost $0.362 on AWS.

3 - What about Spot Instances or Reserved Instances?

There are other questions on my mind, such as what about AWS Spot prices or Reserved Instances? Although Azure offers 6 and 12 month purchase options, they are structured very differently from AWS RIs. With Amazon’s Reserved Instances, users pay an upfront fee for an discounted hourly fee. This is different to Azures two purchase options:

1. Pay Up Front: You can choose to pay a lump sum upfront for 6 or 12 months, and use resources as you need until you reach your spend limit. At the end of the period, anything not used will be forfeited. This option comes with a 22.5% to 32% discount compared to paying on-demand prices.

2. Monthly Commitment: If you do not want to pay up front, you can choose to commit to a minimum monthly payment for 6 or 12 months. If you use below the commitment level, then it is rolled over to the next month until you reach the end of the commitment period. If you exceed the commitment level, then you just pay on-demand. This options comes with a 20% to 29.5% discount compared to paying on-demand prices

This validates our view of a truly multi-cloud marketplace. One in which cloud providers would compete for different parts of your business and will bring different things to the table. The winners in this place will be organisations who are equipped to take advantage of features from multiple cloud providers.

By the way, both Azure and AWS have sponsored the RightScale Compute conference on April 25th and 26th, as well as Google Compute, Rackspace and others. I will be giving a cost budgeting workshop in which I will cover how to optimise on your cloud costs. See you there!

Product Manager at PlanForCloud

Tuesday, 26 March 2013

AWS Consolidated Billing Tips and Tricks

What is AWS Consolidated Billing?
Amazon Consolidated Billing is a feature provided by AWS for a user to group their bills from multiple accounts into one paying account. This enables enterprises to have a single paying account for their finance department, but still get a view of how much each sub-account has incurred:

AWS Consolidated Billing
Consolidated billing is purely an accounting and billing feature, meaning it is not for provisioning or managing resources. Also note that the paying account does not have access to the resources, the dashboard or the data in the sub-accounts. There is no cost for signing up to consolidated billing.

Why use AWS Consolidated Billing?
There are a number of good reasons why enterprises choose to consolidate their accounts into a single paying account:

Monday, 18 March 2013

From PaaS to IaaS: How we migrated off Heroku to RightScale

...and increased our flexibility and outage-proofing in the process

A few weeks ago, we moved PlanForCloud from Heroku, an-all-in-one PaaS, to managing it with RightScale. It was a smooth migration and we're very happy with our shiny new home in RightScale. In terms of cloud infrastructure, we're still using AWS, we've moved from a Ruby PaaS hosted by Heroku in US-East to an IaaS, and we chose US-West Oregon (with RightScale you can deploy to any cloud).

We used Heroku for our first year in operation and found it pretty good, however, we got to a stage where we started to outgrow Heroku. We wanted:

  1. More Flexibility: ability to customise our stack, fine-tune our app's deployment process, get more visibility of our app's performance and be able to select instance types for different parts of our app. For example, we want to use high-CPU instances for our background worker processes.
  2. Outage-Proofing: on Heroku, there is nothing you can do about outages; if Heroku has issues, you'll have issues, and you just have to wait for them to fix it. This is something we faced a few times in which the AWS US-East region had an outage and our entire app went down with it. With RightScale, we can control our deployment to a finer extent and can set up disaster recovery procedures which kick in as soon as any cloud outages happen.
Step By Step From Heroku to RightScale

Tuesday, 12 March 2013

Forecast HP Cloud Costs

PlanForCloud Adds HP Cloud to Multi-Cloud Cost Forecasting

There has not been a single cloud provider who can offer everything for every application and every user. Whether the focus is on cost, security, support or technology, each cloud provider comes to the table with a different offering.

With that, we would like to welcome HP Cloud to the table. HP Cloud servers, storage (both object and block storage), data transfer prices and support plans have been added to PlanForCloud. Here is the server selection page with HP Cloud selected:

PlanForCloud Server Selection Screenshot