Here we are, at the finish line of the Pillars of Successful Cloud Solutions. In the last post of the series, we will cover cloud infrastructure.
When looking at infrastructure for a cloud solution, I break it down into three categories: local infrastructure, connectivity, and cloud infrastructure. Each of these can have a significant impact on the success or failure of a cloud solution deployment. The definitions for each category follow.
Local infrastructure includes laptops, workstations, mobile devices, servers, switches, routers, firewalls, etc.
Connectivity is the path between the local and cloud infrastructures. This includes the connection type (open internet, client VPN, Point to Point VPN, private lines) and bandwidth requirements as well as the hardware on both ends needed to make the connection.
Cloud infrastructure comprises all the physical and virtual components used to run the solution and ensure it is available. This includes networking equipment, physical servers, hypervisor, virtual servers, the physical connection to the internetclient, power, battery backup, cooling systems, monitoring, antivirus, etc.
Below are three private cloud scenarios. Each one lists details of the situation, outlines the issue(s), and provides a solution.
Scenario 1: Local Infrastructure – Local Router capabilities (consumer vs business grade router)
Details: A customer with a three-person accounting department decided to migrate its accounting system to the cloud. In discussions with the cloud provider it is determined a Point to Point VPN will be used to secure the connection between the local and cloud infrastructures. When trying to setup the Point to Point, it is discovered that the local router does not have this capability as it is a consumer grade appliance.
Issue: Consumer grade products often lack business critical components. These missing components are needed when trying to establish a secure connection with a cloud provider.
Solution: Purchase, install, and configure a business class router with built-in Point to Point capabilities, allowing the secure connection to be created between the local network and the Cloud provider.
Scenario 2: Connectivity – Point to Point VPN vs Private Lines
Details: A customer recently migrated its CRM system to the cloud. The ERP system remains on premise. A Point to Point VPN connection has been setup to allow data to be passed back and forth between the ERP and CRM systems securely. Integrations run at different times and priority levels. Inventory, new customers, and sales orders are synced real time (critical), invoices and customer updates are synced every hour. The real time integrations are not performing as expected or consistently, which is creating fulfillment issues.
Issue: After some analysis, latency of the Point to Point VPN connection (Internet) is determined to be the cause. Routers at both ends are operating correctly, it’s the hops in between that are causing the problem.
Solution: Implementing a private line (MPLS, Ethernet, etc.) between the local network and the Cloud provider removes the latency created by the Internet.
Scenario 3: Cloud infrastructure – Adding applicationsusers to existing cloud environment without adding resources.
Details: A customer migrated its production ERP system to the cloud eight months ago. The contracted cloud infrastructure was designed to support 20 Remote Desktop Connection users running the ERP system, MS Office suite, and a standalone reporting solution. After a few initial bumps, the environment has been working very well for the ERP system users. The marketing department has successfully lobbied to have one of its applications migrated to the same cloud. Those 15 users connect to this application via Remote Desktop Connection, just like the ERP system users and had the contents of the local network shared drive uploaded to the cloud for easy access. Immediately after go-live of the marketing department application, the cloud environment began experiencing performance issues. The ERP team noticed the overall response time decrease, the marketing team stated the on premise version of the application was three to four times faster than the cloud deployment and opening files on the cloud server from a local workstation took 30 to 45 seconds longer compared to when the files were local.
Issue: The cloud environment was scoped to handle specific applications with a limited number of users. The database server, Remote Desktop server, and file server were sized appropriately for this initial deployment. When the new application, additional users and local files were added, required resources were not allocated to the cloud environment. Bandwidth usage increase due to a large number of files housed on the cloud being opened by locally installed MS Office applications.
Solution: Increase existing server resources (database server processor, disk and memory, cloud network file share disk space) along with adding another Remote Desktop server. Instruct the marketing users to open their MS Office files on the Remote Desktop server instead of their local computers.
In each case, changes to a different infrastructure category provided the resolution to the individual problem. Understanding not just the high level overview of the cloud solution but its various infrastructure components and how they interact with one another will expedite an answer.
The ideal situation, of course, is to avoid this by having a deployment plan which has taken all of it into consideration. This is where a trusted Cloud Advisor can provide value. Looking at the initial deployment or assessing the impact of additional applications/functions on an existing cloud solution with a focus on all three infrastructure categories can mitigate these risks.
Other Blog entries in the series.
If you are interested in learning more about RSM’s Cloud Computing Rapid Assessment, check out this link.
For more information on RSM’s other offerings please check out our website.
You can also contact RSM’s technology consulting professionals at 800.274.3978 or email us.