This is a continuation of the previous cloud deployment article where I created architectural questions that enable a consultant to understand what products are used to support a corporation's top 10 critical applications. Once these product lists are created, it is much easier to map private or public cloud products that can support these same applications.
These supporting cloud products are a cloud model for a given customer. A cloud provider wants the various corporate applications to share various cloud products while not sacrificing performance and/or disaster recovery. All cloud model permutations should be tested for disaster recovery so that the customer is sure that this can be done in a realistic and successful manner. As I stated in a previous article, it is critical to deploy products that don't support too many permutations.
For example, one could treat a product like a car model and only have pre-configured 'Gold, Silver, and Bronze' engine configurations. This is critical because it reduces the number of testing configurations. The initial suite of product configuration tests can quickly show that the customer's Service Level Agreement is met.
[Related: Five cloud security trends experts see for 2011]
Here are the infrastructure, platform, and supporting application questions used to support the 10 critical business applications. Some of the questions reveal reliance on one vendor only. Others show a disparity in the release/version of operating systems, embedded hardware model, and embedded release/version of products. Other questions show reliance on critical support products like LDAP directories, Domain Name System products, and key management and encryption products.
So let's look at some key cloud-related questions that do not address storage and networking products addressed in the previous article. This list is not comprehensive but it is a good starting point.
Infrastructure Architecture
* Which virtual machines are used to support your critical application?
* What pool of servers do the virtual machines run on?
* What network equipment integrates with and supports the network policy associated with virtual machines?
* Is more than one storage subsystem pointed to by the virtualization system, thereby leading to data replication?
Platform Architecture
* Which servers are dedicated to serving web servers, databases, and application servers?
* Which vendors are used for the various servers?
* What operating systems and versions are running on the servers?
* Which database vendor is used?
* What database release and version is used?
Web Server
* Which web server software vendor are you using?
* What features are critical to the software functioning correctly?
* How does the web server protect application data?
* Which DNS product is used to route internet traffic to the appropriate web servers?
* What release and version of the DNS product is used?
Supporting Products
* What infrastructure management tools are used?
* What network management tools are used?,/li.
* Does it monitor routers, firewalls, load balancers, web proxies?
* What storage management tools are used?
* Does is monitor storage subsystems and storage switches?
* What PKI and encryption servers are used to encrypt data at rest and manage certificates?
* What management tools are used to manage the web servers?
* What LDAP directory is used to perform authentication and authorization functions?
* What tool is used to perform backup (disk to disk or disk to tape)?
* What tool is used to manage corporate email?
In summary, a proper diagnostic reveals the types of products needed to support a client's top 10 corporate applications. The diagnostic enables a service provider to properly map its cloud product set so it can replace and/or augment the current infrastructure. That solution set can be tested to make sure it meets SLAs and disaster recover needs. It is important to note that the cloud layer diagnostics do not address the many supporting applications that are necessary to support the top 10 critical applications. These applications also need careful analysis and must be properly deployed in the cloud solution.