Objectives:

1. Configure GoGrid private network connectivity on Windows 2008 Server and test connectivity to Cacti server
2. Configure and start Microsoft SNMP service on your Windows 2008 Server
3. Add new Cacti device
4. Create graphs to log Local Connection and Local Connection 2 bandwidth and errors, Logged in Users, and server processes

Configure GoGrid private network connectivity on Windows 2008 Server and test connectivity to Cacti server

Below we see that we have a server (“Web2”) deployed on GoGrid with a public IP. Let’s log into this server and configure the private network with a private IP from the same subnet of the Cacti Monitor server. As I described in my previous post – I am using the prescribed private IP subnet from my GoGrid portal, contained under the List tab and then under Network – Private Network.

Once logged into the Windows 2008 server (“Web2”), I go to the Network and Sharing Center which is found by first going to the Start button. From here I need to open up Local Area Connection 2. This is the private network interface that plugs into your own private VLAN on GoGrid. I enter the “Properties” button and then open up “Internet Protocol Version 4 (TCP/IPv4)”. By default GoGrid will enable DHCP for this private network interface. If you have a DHCP server, your server can receive private IP addresses upon initial boot up. Perhaps I will cover this in a different post. For now, we need to configure the interface with a static IP available from our example private subnet 10.129.151.0/24. The subnet will be randomly generated for each customer account. I show you how to find this in my previous post, “How to Monitor Your Ubuntu Server on GoGrid in 6 Steps Using Cacti 0.8.7g.” Following the nomenclature – keeping the last octet of the 32bit address space the same as the public to make IP address association easier – I give this system the private IP address 10.129.151.110. The .110 matches up with the last 3 digits of the public IP address 173.204.58.110. This isn’t necessary, but helps you identify systems and know their private IP address based upon their public address and vice versa. See the image below of me assigning the private IP address.

Now, open up a command prompt >Go to Start, then Run – type in cmd and hit “enter”. I now test connectivity to the 10.129.151.0/24 network by pinging the gateway IP – 10.129.151.1. I get a response, so I can move on. If you don’t get a response, double check the private IP address you used and make sure the Subnet mask is correct. You can also test connectivity by pinging the private IP address of your Cacti Monitoring server. Remember, your 10.x.x.x/24 network will be different than my example subnet. Please check your portal for what you should be using.

Configure and start Microsoft SNMP service on your Windows 2008 Server

Check to see if you have the SNMP Service running already on your Windows system. By default, this server role in Windows Server 2008 on GoGrid isn’t installed and running. Go to Start -> Administrative Tools -> Services. SNMP Trap may be there, but we also need SNMP Service. To install SNMP Service go to Start -> Administrative Tools -> Server Manager -> Select Features -> Add Features. From here find and check SNMP Services (which covers SNMP Service and SNMP WMI Provider). See image below.

Click Next and this will begin the install process of this feature on Windows.

Once the SNMP Service is installed, we can go to the Services page to find and configure the SNMP Service with the appropriate community and host IP address to accept SNMP calls from.

First go to the Services page – Start -> Administrative Tools -> Services or from Run type services.msc

Find the local service – SNMP Service and right click it and go to properties. From here you need to give the service the community string that you will set on the Cacti server and the private IP address of your Cacti Monitor server. This is under Accepted Community Names and Accept SNMP packets from these hosts. See image below:

Add new Cacti device

This step is same “create a device” step that I outlined in my previous post – except the details of the host will be different.

The IP, hostname and template used in the screen shot below represent my example Windows 2008 server named “Web2.” I chose the Host Template Windows 2000/XP Host –> selected SNMP v2 -> put in the community string I chose, and clicked “create.”

You should see the SNMP information for the host quickly appears near the upper left portion of the screen. If you see an error here, you will need to check your private network connectivity between the two servers and check the SNMP Service configuration on the Windows 2008 server.

With the new device in place on Cacti, we can now create the graphs.

Create graphs to log Local Connection and Local Connection 2 bandwidth and errors, Logged in Users, and server processes

From the device we just created, go to Create Graphs for this Host.

From this page we want to add a check to the following graph templates seen below in the image:

1. Logged in Users
2. Processes
3. Local Connection
4. Local Connection 2

Once you have done this, click Create at the bottom of the page.

We will do this again for the In/Out Errors/Discarded Packets option next.

1. Change the “Select a graph type:” near the bottom of the Create Graph page to In/Out Errors/Discarded Packets
2. Next check box the “Local Area Connection” and “Local Area Connection 2”
3. Finally click the Create button at the bottom.

After about 5 minutes, the graph icons will be available and your data will then begin to accumulate for your viewing.

I hope this blog series was helpful for you. The GoGrid Cacti Monitor – Community GSI is a great server-based application that can easily be deployed to a GoGrid virtual server, and configured to communicate via SNMP with your servers on the GoGrid network. The information gathered will give you real-time and historical interface bandwidth, server performance, and other important system level information.

Be sure to check out other Community or Partner Server Images available on GoGrid. The GoGrid Exchange has many pre-configured software solutions that can be deployed to your GoGrid architecture in a matter of minutes.

The main principle behind Infrastructure-as-a-Service (IaaS) is to develop a simpler process to address your IT needs and ensure that it takes less time to plan, manage and deploy – saving you time and money in the end. The question isn’t, “Can cloud computing save you time or money?” because we all know the answer to that. Yes. Of course it can! The real question is, “How much time and money can you save with cloud computing?”

In order to find that answer, you need to have a clear understanding of the benefits of cloud computing and where the real savings come into play. I go over this in great detail in my white paper, Skydiving Through The Clouds. Below are some of the ways companies benefit from transitioning to cloud based infrastructure.

• Economies of Scale – underutilization and misappropriated hardware become a thing of the past. The cloud enables compute, storage and RAM resources to be more efficiently managed, provisioned and scaled, all within a shared environment.
• Cost Control – knowing exactly what you are using and how much that usage will cost you allows for a much stricter control over expenditures.
• “Disposable IT” – companies have the ability to create, use and destroy infrastructure based on their business needs.
• Burstable Workloads– scaling based on demand allows organizations to efficiently use infrastructure in a timely basis.
• Avoiding Capital Expenditure – with cloud computing, infrastructure becomes an operating expense with no need to amortize hardware costs over time.

John Keagy, GoGrid’s Executive Chair and Founder, also jumped in on the conversation a couple weeks ago where he shared what is fact and what is fiction regarding the economics of IaaS in an article entitled The Actual Truth About The Economics of Cloud Computing.

Some cloud providers, like GoGrid, also make very easy to determine what your total operating expenditure will be with cloud hosting calculators. Our tool gives estimated monthly costs based on your infrastructure needs and usage. Remember, GoGrid is “Complex Infrastructure Made Easy™” so we can definitely help you plan out your cloud computing strategy and rollout in a way that is understandable and doable.

Now, I’d like to hear from you. If you are already using cloud infrastructure, how much time and money have you saved since the switch? How are you using the cloud to make your IT-rollouts more cost effective? If you’re still managing your own infrastructure, how much do you estimate you could save by switching to a cloud hosted solution?

Share your answers in the comment section below.

Look forward to seeing your answers!

Today (March 11, 2011), we released our March 2011 update (code name “Long Beach”). The major features and updates are listed below:

• Enhancement to the MyGSI Feature
• Usage Report
• Billing Improvements
• Ubuntu 10.04 Cloud Server Support
• Debian 5.0 Cloud Server Support

Please read on for details on each of the items listed above.

MyGSI Updates

About a year ago, we introduced a new feature called MyGSIs – customized GoGrid Server Images – and how easy it was to create and save a Server Image. Server images are helpful for horizontal scaling, re-imaging of servers, disaster recovery, or setting up failover environments. Until this update, we had a pretty straight forward 3-step process: 1) create a Sandbox Image, 2) configure your server, prep it, and run scripts, and 3) save the server to cloud storage.

Now, we have consolidated this process into ONE step. Simply select the existing server you want to save and click the Save icon. That’s it! There is no need to create a Sandbox Image, run preparation scripts or anything else. And, the best thing is, any existing server can be saved as a server image.

The Save button is also in the List view:

Once you choose the server that you want to save and click the Save icon or link, you will get the following prompt.

Note that if your server has more than 50 GB of data on it, the saving process will fail and you will need to reduce the amount of data on your server.

Simply give your server image a name, description, and indicate whether it is a Web or Database server.

After this, you server will be gracefully shut down, and the image saving process will start. You will see your server’s status light turn to amber:

And there will be an entry in the Jobs tab that not only states that the action has started, but also after a few seconds, displays the expected time for the server to be fully saved to cloud storage. If there is no “Estimated Job Duration,” just wait a minute and refresh the Jobs tab.

Your server will come back on line shortly.

Once the server image is created, you have the ability to instantiate new servers from it, or share it with the GoGrid community as a Community GSI (CGSI).

Lastly, the MyGSI process is backwards compatible, meaning that if you had created a Sandbox Server previously and it still appears within your GoGrid account, you can still convert it into a MyGSI using the same older process. Also, any MyGSIs created using the legacy process will still be available to create new server instances. However, the creation of Sandbox Images has been removed as of this release, and replaced with the much easier process described above.

Be sure to read the MyGSI User Documentation that can be found on the GoGrid Wiki.

USAGE REPORT

Included with the “Long Beach” update is a new feature to help you get a better view into the activity within your GoGrid account. There is now a new tab called “Usage” which is under the My Account section. This new report is in beta for the next 90 days.

The Usage Report overview is broken up into two sections: Metered Services and Fixed Services. Metered services are things like RAM, Storage, and Outbound Transfer. Fixed services are for items that recur regularly during a month’s time, things like monthly licenses or IP addresses allocated to your account. Metered services are more variable in nature.

The Usage Report shows a quick snapshot of Metered and Fixed Services for the current or previous billing period. NOTE:  The historical data in this report will not be backfilled; therefore the previous billing period will not be viewable as of this release. Also, the Usage data collection officially starts on March 11th, 2011. This means that the current billing period report may be inaccurate for up to 30 days for some customers. Once a full billing cycle passes, the Usage Report will show accurate data.

Under the Usage Menu, you can see a link to “Download usage report”. After selecting the billing period from the drop-down menu, click the link to download all of the associated usage data for that period and it will be downloaded to a CSV file.

The CSV file contains the usage statistics for the following objects:

• Cloud Servers – # of GB Hours
• Outbound Transfer – # of GBs
• Licenses – # of instances of a particular license
• Cloud Storage – maximum GBs stored for that day
• Load Balancing – # of instances of a load balancer
• Public IPs – total # of IP addresses allocated to the account
• Operating Systems – # of instances of a particular Operating System in use

The fields reported are listed below:

• Account Name
• Customer ID
• Report Start date/time
• Report End date/time
• Object Type (see above)
• Name of the Object
• Description of the Object
• Data Center
• Usage Start date/time
• Usage End date/time
• Quantity
• Units

A quick tip, be sure to format the “reportStart,” “reportEnd,” “usageStart” and “usageEnd” to mm/dd/yy hh:mm to see the full details with the date and time.

For additional details, visit the GoGrid Customer Portal Guide that discusses the new Usage feature.

Billing Improvements

One of the advantages of utilizing GoGrid’s Cloud Infrastructure Services is the ability to choose a Pay-As-You-Go plan or a Volume pricing plan. We are allotting more focus to the “billing experience.” In the past, our focus was on the delivery of the services and after engaging with our customers, we realize that some time and care was needed for the overall experience. To that end, we have kicked off a complete overhaul of the billing process to provide you with not only a better experience but also with more insight into your account usage.

This first iteration of Billing Improvements consists of the following changes:

1. Monthly Invoices
2. Account Plan Changes
3. Usage

The details of each of these items are listed below.

Monthly Invoices

To make the billing and invoicing process easier, we are removing daily invoicing. From this point forward, customers will now receive monthly invoices on their monthly bill date. The monthly bill date is the date on which a customer opened their account with GoGrid.

There are a few exceptions in which additional invoices may be generated:

• Fixed Services – if you are using a fixed service like dedicated servers, VPN, or hardware firewalls, you will receive an immediate, one-time invoice. The charge(s) for this service(s) will be pro-rated through the next bill date. However, the next monthly invoice will include a full charge for the next month’s fixed service and is paid in advance.
• Manual invoices – if there are manual invoices generated off-cycle.

NOTE: There may be a bit of a transitional period during March 2011 because of the switch from daily to monthly invoicing. If you have any questions, please contact your account manager.

Account Plan Changes

We have updated the process of making changes to your GoGrid account plan and have a clear trail of the changes you have made and when they were made. We have also added email notifications to the process so that you get two email confirmations: one when you request a change and one when the change is implemented.

Just as you could in the past, select either the RAM or Data Transfer Plan and then choose whether you want your plan to be Monthly or Annual (you get a better discount if you choose an Annual plan):

Confirm the plan you want and accept the terms of service.

Once you make your selection, there will be an indication of the plan change within the Plan Selection page (in red) and your current or active plan will appear grayed out:

Also, within the Jobs tab, you will see an entry confirming your plan change.

You may upgrade or downgrade your various plans as you see fit. Downgrades take place on the customer’s next bill date. Plan Upgrades take place retroactively back to the beginning of the current billing period.

The process for Upgrades is a bit more complicated. On the next bill date, customers are credited the last month’s lower plan charge and are charged for the new plan for the previous month and in advance for the next month.

Sometimes examples are a bit easier to understand:

In the case above, you can see how the charges for the lower plan are reversed out and the new plan is retroactively applied.

In most cases, people upgrade plans during a current month because they are going into an overage situation or in anticipation of incurring overage charges.

Ubuntu 10.04 Cloud Server Support

With this release, we are pleased to announce that Ubuntu 10.04 32- and 64-bit Cloud Server Operating Systems are now available to complement Ubuntu within the Dedicated Server Operating System choices. This is the LTS – Long Term Support – version of Ubuntu.

Debian 5.0 Cloud Server Support

Also with this release, we have included Debian 5.0 32- and 64-bit Cloud Server Operating Systems to complement Debian within the Dedicated Server Operating System choices.

More to Come

The “Long Beach” release is the kick-off of a new agile development process here at GoGrid. We will be aiming to have regular monthly releases from this day forward so stay tuned for more exciting news from GoGrid. Be sure to subscribe to the GoGrid Blog Feed to get regular updates.

Before breaking out the checkbook there are several things to consider internally and questions you should ask your potential cloud provider to make sure you are getting the best solution for your business.

Internal Question for your Company

Cloud infrastructure comes in many different shapes and sizes. Having a clear understanding of how your organization will be using this infrastructure will help narrow down which providers will have solutions that fit your need. Here are some questions to ask yourself before approaching vendors:

1. What appliances (e.g., servers, load balancers, firewalls, networks, etc.) need to be supported?Every infrastructure topology is unique, just like your business. In order to create the best infrastructure solution, you need to consider what virtual and physical appliances are required or optional to your cloud solution.
2. What security requirements do you have? Is a shared environment acceptable or does your cloud infrastructure need to be single-tenant?Cloud computing can come in multi-tenant and single-tenant options. Single-tenant infrastructures are 100% dedicated to your company, but they do come with an additional price. And within public clouds, you can segment off a portion of your infrastructure (e.g., dedicated or colocated environments) that is single-tenant (see Hybrid Hosting).
3. What are some of the different use cases that need to be supported?Understanding how much computing power your company needs, when your company needs it (seasonal) and which departments will use cloud infrastructure is useful in selecting cloud infrastructure packages. Your first step is to clearly define these business models and use cases so that a custom solution can be created to meet your needs.
4. Who will manage the implementation?Many companies are happy with a do-it-yourself option in terms of creating a cloud infrastructure for their company. But there are plenty of time and cost constraints that come with these types of implementation. You can also choose a cloud provider that consults with you to understand and help you implement your environments in the most cost and time-effective way. And once the implementation is live, who will be there to support it? Your internal team or the cloud provider? (Be sure to look at SLAs as well.)
5. Are there regulatory requirements to consider?If there are compliance, regulatory requirements or unique/customer hardware considerations that are critical to your cloud implementation, be sure to map those out ahead of time.

Questions to Ask your Cloud Provider

Now that you know exactly what YOU are looking for, it’s time to find a cloud provider that builds solutions tailored to your needs. Simply sharing your business’s needs to an infrastructure-as-a-service vendor will get you recommendations and solutions, but they may come with some unexpected surprises. It’s a best practice to know the vendors process and procedure to make the purchase and implementation as smooth as possible. Here are some things to consider:

1. Is it a scalable solution from both a company-installation standpoint and an end-user usage standpoint?You need to be sure that whatever cloud provider (or partner) you choose can grow with your business, otherwise you do not get the advantages of cloud computing. Is your cloud partner large enough to scale with your needs? And can users of your cloud account scale the infrastructure as well on-demand and just pay for what they use? (Beware of the “false cloud” where you have to shell out money in terms of capital expenditure – buying hardware to power your cloud is NOT cloud computing.)
   
2. Are they simply a reseller of someone else’s technology or are they an innovator in the space?Several of the big players in the cloud space are simply resellers of Infrastructure as a Service. Or, they have an older infrastructure solution and they are simply slapping the word “cloud” in front of it (this is known as “cloud washing.”) If your needs are simple, they can be great partners. If your computing needs are more complex, it makes sense to partner with a company who is considered an innovator.
3. Do they require capital expenditures?Believe it or not, several cloud providers will require capital expenditure and on-site hardware. Again, it is not a cloud solution if you have to buy hardware to power it and then manage that infrastructure yourself (“false cloud”). Having to buy physical hardware to power and grow your company’s “cloud solution” really just causes more work for your IT staff and isn’t cloud computing.
4. Do they provide technical expertise when creating the implementation?The hardest part of cloud computing is the implementation or migration process. Make sure you have the support you need during the transition. Will your cloud partner help you overcome any technical hurdles or offer best practices?
5. Is their offering based on industry standards or are you required to architect your infrastructure design to meet THEIR requirements?This is big. Some cloud providers require you to re-architect your infrastructure design to meet the requirements of their stack. This can often lead to more work for your company.

We hope that these questions will help you find the right solution for your company when you are looking to acquire cloud infrastructure. We’ve identified even more considerations in our white paper, “Skydiving Through the Clouds”.

Download “Skydiving Through the Clouds”