Last week, we held a web session introducing ParkMyCloud for Microsoft Azure. We’re excited to open up the ParkMyCloud platform to Azure customers, so you can get the same savings that AWS customers have been enjoying for the past year and a half.
Watch the video here, and use the guide below to skip ahead to the parts of the session that interest you:
00:28What is ParkMyCloud?
WHAT: Simple, single-purpose SaaS tool.
HOW: Automatically schedule on/off times for idle servers.
WHY: Optimize cloud services spending.
ROI: Save 60% or more; 6 week payback.
01:13How ParkMyCloud Works
•Discover & Manage cloud computing resources
•Analyze & Recommend resources to ‘Park’
•Policies automatically schedule resources for off/on
01:48ParkMyCloud vs. AWS cost savings options
02:15How much has ParkMyCloud saved our customers?
02:39The Azure World
Your Azure Account (Active Directory Tenant ID)
Your Azure Subscription (Subscription ID)
Your Application (Application ID)
Service Principal for Your Application
Limited Active Directory Role for Service Principal
Your Azure Cloud Resource Groups and Resources
Azure limited access role credential is analogous to an AWS IAM Role + policies
Requires a lot more information:
Subscription ID ~ analogous to an AWS account
Tenant ID ~ the ID of our Azure AD instance
App ID ~ the ID of the ParkMyCloud App in AD
Password (a.k.a., Client Secret) ~ You set this
04:39Four Azure CLI Approaches to creating Azure credentials
Windows Powershell – manual or scripted
Unix azure-cli – manual or scripted
05:42 ParkMyCloud Demo
05:51 PMC Dashboard
06:05 Adding an Azure credential to ParkMyCloud
07:29 Walkthrough of the ParkMyCloud dashboard
08:27 How to attach a schedule to an instance
09:18 How to create a custom schedule
10:51 Your savings projections in ParkMyCloud
11:51 How to see information about your individual instances
12:34 Teams and Roles in ParkMyCloud – containers for organizing users and resources with role-based access control (RBAC)
14:22 Logical Groups – the ParkMyCloud construct for organizing resources for group scheduling and sequencing
16:51 Policy engine – apply schedules in an automated fashion
17:02 Never Park policy – protect production instances from parking
17:26 Creating a new policy for scheduling
19:44 Always Off Schedule – use to park for the maximum amount of time. Useful when users are across time zones.
22:09 Audience Question – using the policy engine for sorting to teams, “snooze only” for schedule enforcement, and others.
23:37 Actual Savings number
24:04 Quick filters for viewing the dashboard
24:32 Recommendations – how to edit and add recommendations, and parking resources that are recommended to park.
25:00 How to download reports in ParkMyCloud
25:12 Audit Log
25:33 Pricing – by instance count
26:45 Is there any functionality for AWS that doesn’t translate to Azure?
28:59 Do you plan to support other clouds besides AWS and Azure?
29:42 How are the projected savings and actual savings numbers calculated?
Auto-Scheduling for Microsoft Azure Opens Door for Microsoft Azure Customers to Save Millions
January 24, 2017 (Sterling, VA) – ParkMyCloud, the leading enterprise app for optimizing and reducing cloud spend, today announced that it now supports auto-scheduling for Microsoft Azure in addition to Amazon Web Services (AWS). ParkMyCloud launched its “Nest for the cloud” platform in September 2015 to enable AWS customers to automatically turn off idle instances, saving 20-60% on their cloud bills every month. The company has seen rapid customer growth, and customers include companies such as McDonald’s, Sage Software, Neustar, Avid, and Wolters Kluwer.
With Azure, ParkMyCloud significantly broadens its public cloud appeal. AWS has the largest Infrastructure-as-a-Service (IaaS) market share at about 30%, while Azure comes in second at 11%. Adding the second-largest provider means that a greater share of public cloud users can now manage and optimize their resources and get savings on their cloud environments. Additionally, ParkMyCloud now supports not only customers who primarily use one provider or the other, but others who have a multi-cloud environment crossing both AWS and Azure.
451 Research Vice President William Fellows commented on the announcement, saying, “After ParkMyCloud’s work in 2016 deepening their functionality in AWS, supporting Azure helps them target this additional market. The single-purpose tool approach helps ParkMyCloud efficiently address the cloud cost control market and Azure customers should appreciate the new opportunity for simple savings.”
“During 2016, our customers saved over $1.6 million on their cloud bills using our innovative approach to cost savings,” said ParkMyCloud CEO Jay Chapel. “Support for Azure was the top requested feature, so today’s launch will help us drive even bigger growth during 2017 as we become a go-to resource for DevOps and IT users on all the major cloud service providers.”
ParkMyCloud plans to support Google Compute Engine early this year. Following this expansion, ParkMyCloud will broaden its cost-saving offerings to cover databases, storage, and rightsizing.
ParkMyCloud is a simple, single-purpose SaaS tool that enables users to automatically schedule on/off times for their idle cloud computing servers (also known as “parking”). Users save up to 60% on cloud server spend by paying only for the time they actually use to avoid wasted spending. Customers include McDonald’s, Sage Software, Neustar, Avid, Wolters Kluwer and Tristar Medical Group. For more information, visit http://www.parkmycloud.com.
Today I’d like to tell you a story* about how one company learned to control cloud costs. Let me introduce you to Jack.
Jack is the instance slayer.
“The instance slayer?” you ask. “What does that mean? It sounds fun and powerful.”
Powerful? Perhaps. Fun? Not exactly.
You see, Jack is a software engineer at a large financial services company. Unlike some of his co-workers who are more specialized, Jack is, well, a jack-of-all-trades. He does a little in mobile, a little in infrastructure, a little in architecture. He’s busy.
One day, Jack’s boss came to him with a new task.
“We need to take control of our department’s cloud instances to control cloud costs,” his boss said. “I need you to come up with a way to ensure that instances are turned off when they’re not being used.”
So, Jack developed a set of scripts that determined whether everyone’s instances met the new on/off compliance standards. These standards required non-production instances to be “stopped” when not being used. If they were left running when they shouldn’t have been – and a lot were – Jack would first send an email to the owners of the offending instances. But if after the email they still didn’t comply, Jack’s scripts would terminate the instances overnight.
As you can imagine, people got quite annoyed when Jack’s scripts terminated their instances.
“Well, I warned you,” he would say. What could Jack do? This wasn’t his idea. He understood that there was a lot of wasted cloud spend each month, but ultimately this was just a task he was carrying out at the request of his manager.
Nonetheless, Jack made a couple of enemies. They’d call him the Terminator. The Instance Slayer.
Jack didn’t mind the enemies. It was the time out of his day that it took to wrangle them, email them, and clean up their instances that irked him. He had things on his to-do list that he’d much rather spend his time on.
Then, Jack’s teammate Roland introduced ParkMyCloud to the team. A bunch of people tried it out, including Jack.
“This is great,” Jack told Roland. “If I set up simple automated policies to put everyone’s instances on schedules, I won’t need to chase them about turning their resources off when they’re not being used. I’ll have more time for my other projects.”
Everyone else was happy, too. With ParkMyCloud’s simple governance model, Jack and the other administrators could add users and create teams, enabling end-users to apply their own schedules. Plus they got reporting to know if instances weren’t complying. Remember those scripts? Well, Jack didn’t need those anymore.
Jack gets free time: Jack wins. Jack’s teammates get more control over their resources: Jack’s teammates win. Jack’s company gets to control cloud costs by cutting wasted spend on unused cloud time: Jack’s company wins. Win, win, win.
Jack may have stood down as the instance slayer, but his name will live on in legend.
*This is based on a true story! Names changed & company anonymized to protect our customer’s privacy – especially Jack.
The other day, we talked to a prospective ParkMyCloud customer about how to protect his production servers. He had just started a trial of ParkMyCloud, and before he added additional users to his account, he had an important question: How can I keep my production servers safe from my end users accidentally parking them?
It’s a great question! Before you start shutting down your non-production resources for cost savings, it’s a good idea to protect your mission-critical production resources from being parked, which could wreak havoc on your applications — while some resources can be stopped, or “parked”, during off-hours, there are, of course, others that need to run 24×7.
Luckily, with ParkMyCloud’s policy engine, it’s straightforward and easy to protect your production resources. All you need to do is apply a “never park” policy so those resources cannot be scheduled or manually started/stopped.
Only users with the SuperAdmin role can create and manage policies, so if you’re the primary account holder, you don’t have to worry about end users changing these policies once they’re set up. Your production resources will be safe from being parked, so you can start parking and saving away.
Go to the left sidebar and select “Policies” and “Create Policy”.
Name your policy – let’s call it ‘Never Park’
Input the criteria to identify your production resources – usually this will be by name or tag.
Select “Restrict” as the action.
From the Restrict dropdown menu, select “Never Park”. For this option, users can neither attach schedules nor manually start/stop resources.
You can also use policies to assign instances to teams based on their tags or names. If you set up your teams in advance, this is a simple way to automatically control which users have access to which instances. So in this example you could set up a “Production Team” and have all your production instances sort directly to this team. And as an admin you are able to create the permissions for who has access to this team, adding another layer of protection.
Save your policy to protect your production resources from being parked and you’re all set!
There are a few other reasons you may want to use policies on your resources. For example, you can use policies to automatically attach or detach schedules to instances, again based on credential, location, name, type, or tag. So, for example, you could set all of your instances tagged “development” to have the “Up M-F, 8 am – 5 pm” schedule automatically applied.
You can also restrict resources with parking schedules to “snooze only”. That is, end users can only snooze the attached schedule, they cannot edit, detach or change it.
The policy engine is a powerful feature that can help you automate many of the common actions within ParkMyCloud, ensuring that you maximize your cost savings with the least amount of effort.
If you have any additional questions about using policies, please comment below or contact us!
We recently held a web session “5 reasons to turn off your cloud servers when they’re not in use.” Even if you weren’t able to join us, you can check out the recording below – or use the annotations below to skip ahead to the section you’re interested in.
00:51 Overview of presentation
01:18 Reason #1 – Save Money
Compute makes up the largest share of spend for most enterprise cloud users
$10 billion/year in revenue
70% is EC2 ($7 billion)
52% is non-production which could be turned off at night ($3.5 billion)
A simple 12-hour on/off schedule saves $1.75 billion in instance costs
03:00 Comparison of the major ways to save money in AWS
03:07 Reserved instances – pay upfront for a contract for 30-40% savings. Best for production environments.
04:06 Spot instances – open market on spare capacity. Savings are compelling (70-90%) but risk that your instance could go away if the price goes above your bid price.
04:37 Autoscaling groups – scale up and down groups of all types of instances according to your environment needs.
05:00 Turn instances off to save money when you’re not using them.
05:29 Reason #2 – Improve Security
There is nothing more secure than a server that is OFF (This is what security people actually dream about)
There is nothing easier to monitor than servers that are OFF (so long as your NOC folks are NOT surprised)
06:20 Reason #3 – Reduce Environmental Impact
Data centers consume ~2% of all electricity globally (3% of U.S.), growing at 12% YoY.1
80% of power is provided by fossil fuels
Avoiding the need for new data centers reduces carbon emissions
It also makes good business sense:
Cloud providers like AWS improve profitability and increase carbon credits when they can oversubscribe their infrastructure.
They can’t oversubscribe if they are over-provisioned
Therefore, turning servers off when not in use is one sure fire way help oversubscribe
08:31 Reason #4 – Because Werner Says So
AWS CTO Werner Vogels recently addressed an audience of AWS users, saying, “You can turn off your resources when you go home. Typical cost savings are 75%.”
08:51 Reason #5 – Peace of Mind
Knowing you’re saving money, improving security, helping the environment and obeying Werner will help you sleep better at night … who wouldn’t?
The question is how?
You manually turn things off when you leave. In Offices there are automated light switches. There’s a reason, right?
You could create scripts to do this automatically – that is not a cost-effective
09:29 ParkMyCloud – built specifically to turn off your servers when you don’t need them
WHAT: Simple, single- purpose SaaS tool.
HOW: Automatically schedule on/off times for idle servers.
WHY: Optimize cloud services spending.
ROI: Save 60% or more; 6 week payback.
10:17 ParkMyCloud product demo
10:28 Dashboard overview – see all your instances in one place
11:19 Parking recommendations
12:15 Parking schedule interface – create your own schedules
13:51 Snooze parking schedules
15:24 Automated policy engine – create policies to automatically handle scheduling, team assignments
18:16 How did you connect to AWS and find your instances?
19:18 How do you figure that you can save 65% on instances that you turn off nights and weekends?
19:52 Can you compare turning instances off to Reserved Instances?
22:12 Do the energy savings you described for Amazon apply to other cloud providers?
24:05 Why Dale doesn’t like scripting on/off times
We welcome additional questions in the comments below, and hope to see you at our next web session!