Change is inevitable for an enterprise application life cycle. The change for an enterprise application comes in various forms and the most common words used for the application changes are migration, transition, transformation, upgrades and consolidation. The diagram below creates a layered architecture for an enterprise application changes named as Enterprise App Migration Layers. The Enterprise App Migration Layered Architecture provides a cross industry standard approaches for application migration needs.

It is leading practice to limit number of layers changing for a given migration event to keep the migration simple as much as possible. In contrary, you get economies of scope benefits to combine the changes in multiple layers. Therefore, selecting changes for a migration event is an art and science that needs to combine with the capabilities, refresh requirements (usually driven around app platform components end of life timelines), timeline objectives and cross functional resources availability.

• 7. App Consolidation - App Merger & Acquisition: When an enterprise have more then two applications serving one functional domain that either applications is capable to serve functional users. In this case, enterprise architect often port less desired application stack to the most desired application stack. On the other hand, if a functional domain is matured and common in industry that has an economical Cloud Computing Software as a Service (SaaS) provider(s) that meets your functional, performance and security requirements. An enterprise architect may consider to migrate to a SaaS provider. It is a leading practice to develop and execute comprehensive testing plan with multiple domains such as features, system, integration, security, performance, backup & disaster recovery and user acceptance. This layer migration usually driven by top down enterprise IT strategy with or without cloud.


• 6. Platform Refresh - App Refresh: You are refreshing application COTS product to latest version or adding new components for new features either on existing or latest versions of underlying technology stack. The migration is manual in nature but one may consider build automation. It is a leading practice to develop and execute testing plan in domains such as features, system, integration, and user acceptance.


• 5. Platform Consolidation - App Remediation: You are retaining application COTS product version and functionality but porting underlying technology stack to desired vendors to comply with enterprise architecture mandates such as DB2 to Oracle, WebLogic to Tomcat, IIS to Apache etc. The migration is manual in nature but one may consider build automation. The most common type of migration when an enterprise is looking to eliminate expensive technology components from the enterprise technology portfolio or migrating to Public Cloud Computing - Infrastructure as a Service (IaaS) provider. It is a leading practice to develop and execute testing plan in domains such as features, system, performance, integration, and user acceptance.


• 4. Middleware Refresh - App Lift-n-Shift: You are retaining application COTS product version and functionality but upgrading the underlying middleware technologies to latest version such as Oracle DB, WebLogic AS and Apache WS. The migration is manual in nature but one may consider build automation. It is a leading practice to develop and execute testing plan in domains such as system, security, performance, integration, and user acceptance.


• 3. OS Refresh - Middleware Lift-n-Shift: You are upgrading versions of underlying operating system or migrating to an standardize operating system image of existing version. One may consider tools like AppZero or manual transformation. The most common migration type when an enterprise is changing and/or outsourcing to managed hosting provider. It is a leading practice to develop and execute testing plan in domains such as system, security, integration, and user acceptance.


• 2. Infrastructure Refresh - Logical Lift-n-Shift: You are refreshing server hardware platform only. It is a leading practice to leverage tools to perform Logical Lift-n-Shift migration. The industry leading tools for such migration are PlateSpin Migrate, Rackware, VMWare Tools, and Data Gardens. It is common to hear the terms like P2P, P2V, V2P and V2V migrations for this kind of exercise. The most common migration type when you are migrating to Cloud Computing - Infrastructure as a Service (IaaS) provider.It is a leading practice to develop and execute testing plan in domains such as system, security, performance, integration, and user acceptance.


• 1. Tech Center Refresh - Physical Lift-n-Shift: You are either moving from on-premise data center to colocation or changing colocation provider only. It is a leading practice to develop and execute testing plan in domains such as integration, and user acceptance.

Cloud Readiness of an application system deployment in IaaS service model is dependent on the following factors:

• Technology: Operating systems and middleware technologies are capable to run on x86 server architecture platform also known as x86 server platform. The most commonly used operating systems over x86 server platforms are Windows and Linux operating systems and widely available as IaaS from several vendors such as Savvis, Rackspace, Terremark, and Amazon.
• Implementation: We often find that our implementation is not compatible to run on x86 server platform and have direct dependency on the server hardware platform. In order to be cloud ready, it needs to be on a x86 server platform with no direct dependency on the server hardware platform.
• License and Support: This one tightly aligns with the products used for implementing the application. Product vendor needs to support and economical license options for an application system deployment in a public and private cloud deployment. You may find that to overcome of license constraint people produce a dedicated cluster for certain technology component such as Oracle Database.
• Scale and Performance: The public cloud vendors normally will not guarantee the performance for disk and network IO. You will find dedicated CPU and Memory with System Availability commitments from the public cloud vendors. If an organization got its own scale to operate and has strict performance requirements around disk and network IO and doesn’t need to scale (beyond pre-defined limits) without a lead time are implementing a managed private cloud.
• Security Mandates: The most difficult security mandates to meet by a public cloud vendor is that my data needs to stay physically separated from others. In that case, you are required to implement a private cloud. The security mandates that requires data to be encrypted over wire and rest and who can access can be met by a public cloud vendor offering using software based controls. We may find that organization go for a hybrid model where you hand pick what to share (software controls meets your security mandates) and what you won’t (you will need to have hardware controls in place to meet your security mandates). Cloud provider such as Savvis specializes in the creating the Hybrid System Deployment to meet the Enterprise Cloud Computing (ECC) needs.

First of all, let's understand Enterprise Service Oriented Architecture (SOA). In nutshell, the service catalog for an enterprise SOA is built for re-usability by multiple internal and external consumers, not for mere system integration for a single consumer.

The Enterprise SOA adoption strategy pillars are Platform Value, Platform Commitment, and Platform Activities. Using these pillars, one can develop SOA adoption strategy and plan that suit their organization values and culture:

• Platform Values: Ability to deliver quality products faster and cheaper. For business, it is all about economies of scope and scale. For IT, it is plug and plays development as much as possible that is not re-inventing the wheels. In other words, re-usability.
• Platform Commitments: To win application owners support and confidence the platform needs to make commitments to be available, scalable and backward compatible.
• Platform Activities: Establish SOA collaboration, not just governance. Create continuous education for business and IT. Catch at the conception to onboard new applications and proactively refine service catalog to onboard new and existing applications.

Data Center Migration Project (DCMP) are usually expensive and considered risky. I personally believe and proved several times that it is an opportunity to improve infrastructure deployment and exercise standardization & tech refresh as economies of scope than just refreshing compute, storage and network platform. We mostly find that most organizations just limit the scope of DCMP to refresh compute, storage and network platform and miss the huge opportunity to kill multiple birds in one shot.

Data Center Migration Project Strategy pillars are Migration Waves, Migration Methods and Migration Validations. Using these pillars, one can develop DCMP strategy and plan that's suits their organization values, culture and applications portfolio:

• Migration Waves: It is critical to divide your application portfolio into multiple waves based on business, application and team needs. At the minimum, you are at least planning two Migration Waves for a DCMP. Usually the first wave that is also known as Wave Zero consist of standing up core infrastructures of compute, storage and network with core system services such as DNS / GTMs, AD / Single-Sign-On, Firewalls Stage One, L4 Load Balancers / LTMs, Mail Relays, Monitoring, Central Logging, Outbound Proxy, Backup and Recovery, and Misc Services (VPN, SVN, Terminal Servers, Jump Servers). These core systems services usually could co-exists with current production either independently or in-collaboration.
• Migration Methods: Selecting a method is key that is dependent on availability needs of business function, application design, tools capabilities and skills of the team. The commonly known methods are first green field build, second clone the currently running production server from backup and third clone the currently running production server fresh. Each migration method has costs, benefits and risk associated with it.
• Migration Validations:I usually refer this as minimum verification plan that needs to created carefully. We all do great jobs with specific verticals and the migration issues falls in cross-boundary. Identify and develop cross-boundary validation and have an enterprise solutions architect (who understands applications and infrastructures components) available to provide the technical oversight for solving cross-boundary issues faster and cheaper.

Fast Public DNS Servers are available free for internet users from trusted vendors like Google, Verizon, DNSAdvantage, OpenDNS and Norton. You also have choice to leave default DNS servers provided by the Internet Service Provider you are currently connected to go to the internet aka IPS's DNS.

You have advantage in configuring your personal laptop using the Fast Public DNS Servers than ISP's DNS by getting the speed of name resolution and protection from the man in the middle and domain phishing attacks foster by a free WiFI and Internet Service Providers.



On MS Windows: Open network connections. Right Click Properties. Open the TCP/IP Properties. Click on Use the Following DNS Server Addresses. Enter your choice from the Free DNS Service Provider table on left. It is recommend to pick the Public DNS Servers from the different service providers for high availability user experience.

In Enterprise setup like coroporate LAN/WAN one must use primary DNS provider from Corporate Intranet that usually starts at 10.x to resolve the corporate intranet DNS site domains and you may select secondary DNS server from a Publlic DNS service provider.

IP Address based User Persistence and/or Stickiness works for the users whose public IP address is fixed during user session and their ISP is not using stateless forward proxies to serve the http/https request to the outside world.

Cookies based User Persistence and/or Stickiness works in all the cases where even an ISP provider like AOL who uses stateless forward proxies to serve the http/https request for their customer for better user experience. The implementation of the forward proxies to serve http/https request from enterprise networks is also used not only for better performance and network resource optimization but for better security and monitoring.

Cookie based User Persistence and/or Stickiness will always works even if the stateless forward http/https proxy is used to serve the end users. It recommended that we use User Cookie over IP Address.

What is Server Virtualization? Server Virtualization is all about sharing a hardware either to scale an application or between two or more different application(s). As shown in the diagram, system architects can implement Server Virtualization either by Logical or Physical Virtualization design pattern. In logical server Virtualization hardware and operating system is shared while in physical server Virtualization only hardware is shared.
In real life example to share and utilize a large mansion for logical Virtualization we could allocate an individual bedroom to a person with a common understanding to share the kitchen (Operating System) while for physical Virtualization you invited the building contractor (VM Manager) to create an apartment complex with it's own kitchen (Operating System) for each person.

Why Server Virtualization? The top 3 drivers for implementing Server Virtualization are

• Low System Utilization : Over the last decade, the hardware innovation has left application software components behind. Recent surveys indicated that industry consumes little over 15% computing power available in the data centers. Server Virtualization can you help you boosts the System Utilization.


• High COTS Software License and Support Cost : Today's dynamism in the business worlds demands that we develop the processes like Agile / Scrum to quickly roll out solutions and updates that has created a need for multiple development, test and pre-production environments resulting in high COTS Software License and Support Cost. Logical Virtualization has edge over Physical Virtualization for saving COTS Software License and Support Cost until industry shifts the software pricing models.


• Rising Data Center Cost : Hardware has become economical over the time but the increase in cost for rack space and powering them in a data center environment has become more expansive that is demanding that we look for ways to optimize and consolidate data centers.

Disclaimer

The views expressed in the blog are those of the author and do not represent necessarily the official policy or position of any other agency, organization, employer, or company. Assumptions made in the study are not reflective of the stand of any entity other than the author. Since we are critically-thinking human beings, these views are always subject to change, revision, and rethinking without notice. While reasonable efforts have been made to obtain accurate information, the author makes no warranty, expressed or implied, as to its accuracy.