Where Do Companies Want to Land on the Migration Continuum?
As mentioned, there is no one perfect landing spot on the migration continuum that is right for every company. Along that continuum, there are advantages and disadvantages that have to be considered when companies look to migrate from their current environment.
Let’s take a look at those advantages and disadvantages of the 4 possible landing spots mentioned earlier.
Readily available data:
As companies look to isolate data from processing in order to help make better business decisions, they can consider multiple ways to move the data to a lower cost platform where it is easier to access. ACNielsen in the 1970s, Sperry Univac in the mid 1970s, and Teradata in the mid 1980s were some of the well-known early providers of data warehousing capabilities. In its earliest form data warehousing moved the value of data from the operational business flow to the decision-making process. Data Warehouses in essence are central repositories of data accumulated from multiple sources. They store this data in one single place that can then be used for creating analytical reports for the business community. In terms of the major reasons that organizations look to migrate from the mainframe, the landing spot of readily available data certainly helps satisfy the goal of extensibility.
Data Warehouses provide the ability to use business data to make informed decisions. It allows the evolution to technologies such as predictive analytics whereby the accumulated data can be used as a resource to make better business decisions based on current trends. It is a dynamic view of data that is far different from the static view of many historical mainframe data storage alternatives. It also is a very low risk alternative to moving workload from the mainframe to a lower cost platform, as the areas of data accumulation stay on the mainframe with unchanged business rules.
Unfortunately, in most cases, the data in Data Warehouses is snapshot data. Using OLAP (Online Analytical Processing) it is possible to handle lower volumes of transactions because of the complex aggregation requirements of analyzing that data. OLTP (Online Transaction Processing) has not been as prevalent as OLAP in the history of Data Warehousing, as the impact to processing tends to increase cost instead of decreasing it. In addition, data being moved to a Data Warehouse most often has to go through a data cleansing cycle to ensure that any faulty data is identified and corrected before being fed into the Data Warehouse model.
The Lift-and-Shift model has been the migration model of choice for over 20 years. As corporations have been forced to consider alternatives to mainframe processing, many have seen Lift-and-Shift as the lowest risk alternative. One of the major benefits of the mainframe has always been its low risk factor. The beauty of the way mainframes work is their dependability – they do the same thing the same way every single day for years and there is very little risk that things don’t work the way they are supposed to. So, for organizations to consider moving away from that environment, it is imperative that they find a landing platform that offers that same low level of risk. In terms of the migration continuum, cost is a factor directly addressed by the choice of a Lift-and-Shift model. And that is exactly what Lift-and-Shift offers.
Rather than refactoring code or replacing business rules with a packaged solution, those business rules, the things that in truth differentiate one company from its competitors, stay in place. If the mainframe language is COBOL, those business rules stay exactly the same on a lower cost platform still running COBOL. If users have been used to a certain data entry screen format in the way that they have worked for 20 years, that user interface stays exactly the same after migration.
Although Lift-and-Shift does not directly address the resource availability that many organizations have, it does position them in an environment in which experienced mainframe programmers are using the same development environment and tools that their non-mainframe compatriots are using. That means that the ability to work more closely together alleviates some of the concern of resource availability and can serve as a starting point for some of those experienced mainframe resources to start to develop skills using more current tools.
Just based on the fact that Lift-and-Shift is designed to ensure that things work as they did on the mainframe, that model does not lend itself to benefits like extensibility. What many companies do is make Lift-and-Shift a “step 1” in a multi-step process. Their landing spot on the continuum is at a place where they will see benefits in cost marked along with low risk, and then use the monetary savings to evolutionarily move forward on the continuum. They can invest in newer technologies and languages that help them ensure they are responsive to their ever-changing business needs.
As newer technologies and development tools have evolved, many companies who use mainframes have been concerned that their aging staff has become a bottleneck. They see development off of the mainframe occurring at a faster pace than they see development done on the mainframe and they worry that their tie to some of the older technologies makes them less agile than they need to be. An example might be a traditional retailer with a brick and mortar user environment where people come in and walk the store looking for what they want. In today’s world though, we all know that there are more and more people who don’t shop that way – instead they get onto the mobile device or tablet and do online shopping. So, doing business as companies always have too often leaves them at a competitive disadvantage.
The same could be said for their development environments. Companies need to provide a responsive development and production environment that is responsive to an ever-changing business climate. Finding .Net developers or java developers who can quickly assimilate the newer tools and technologies affords organizations the ability to more easily stay competitive. And as mentioned earlier, more and more of the traditional mainframe developers are retiring and there are not as many avenues to explore as there once were in finding places from which to replace them. So, for companies who worry about resource availability, landing on the Code Refactoring portion of the continuum can make perfect sense for them
While resource availability is one obvious benefit of code refactoring, it is important to understand that development agility is equally important. The rate of new software paradigms only increases every year and working in a platform that easily adapts to those paradigms is ever more important. Being able to leverage existing objects, interface to existing entire application areas like Nuget, and using development environments more attuned to rapid change can make the difference in a company being an industry leader or being an industry afterthought. The real key to code refactoring is to work with a solution set that yields quality code. There are multiple choices in refactoring code – you can manually go through and extract business rules and recode them by hand, you can have automated tools that follow a rigid set of transformation rules, or you can have a combination of the two. The key is to ensure that the code you end up with is of high quality and high maintainability.
Code Refactoring, if done by hand can be very time exhaustive and risky. You rely on human beings ensuring that they don’t overlook or misconstrue existing business rules. And if you use a transformation tool that builds resultant code, in many cases the code is not easily manageable or extensible. Most mainframe languages are procedural in nature and most “newer” languages are object oriented in nature. If not careful, you can end up with code that is a hodge-podge of the two, and which does not give you the real advantage of ending up in an object-oriented model. That being said. Astadia has worked with the leading providers of code refactoring tools to ensure that there are automated processes that can build quality, maintainable and extensible object-oriented code that can be the basis of an organization’s code-set moving forward.
At Astadia, we say the cloud IS the new mainframe. Whether a company is looking to move data so it is more readily available, Lift-and-Shift their applications, or refactor their code, the cloud can be a part of that solution. In terms of the reasons discussed about why companies look to migrate from the mainframe, the Cloud adds directly to those benefits.
Scalability and availability cost savings that the Cloud provides only help to extend the benefits of cost savings that people see when they perform a Lift-and-Shift migration. The Cloud can offer an advantage in terms of resource availability, especially in an IaaS (Infrastructure as a Service) environment where companies are worried about losing their very senior and experienced system programmers who are the lifeblood of keeping the mainframes running. And in terms of extensibility, the Cloud through implementation of
technologies like Containers and Kubernetes allow an environment perfectly suited to extending processing capabilities via more flexible deployment models and interface to scalable data stores.
The Cloud, in its various states, is a proven model that does in fact afford tremendous benefits to organization who move there. Lower costs, flexibility in licensing, ability to lessen the human resource element to maintaining a data center are all proven benefits to a cloud adoption model. Other benefits of a cloud adoption model would include the following:
> Ease of use – the Cloud is designed with simplicity in mind. You can request new services and host your applications using well defined deployment tools.
> Flexibility - You can select from a wide variety of virtual environments where you choose the software and services your application requires. If you find that the environment selections are not adequate, you can simply provision different types of instances or add compute and/or storage on demand.
> Cost-Effectiveness – Cloud usage is billed in a consumption model, where you only pay for the compute and storage resources you use with no upfront commitments and contracts. Alternatively, if you know the minimum compute capacity required during an extended period of time, you can reserve capacity at a significant discount.
> Reliability - With the Cloud, you are taking advantage of its highly redundant, worldwide computing infrastructure that is designed with strong high availability capabilities, which satisfy some of the most stringent enterprise requirements.
> Scalability – the Cloud includes features such as Auto Scaling and Load Balancing which allow your application to scale in or out, if you design/architect this into your solution. The Cloud’s massive compute and storage infrastructure is designed to make resources available when they are needed.
> High Performance – The Cloud offers a wide selection of compute and storage options to satisfy the performance needs of your formerly mainframe-based applications. Compute and storage can be provisioned as they are needed, so if your application is CPU intensive you can have a larger CPU/IO ratio and vice versa.
> Security – The Cloud provides multiple security capabilities and services to improve privacy and network response
With the Cloud, organizations have the choice of having public, private or hybrid deployment choices. For some companies in the early days of cloud deployment, the requirement to not allow data outside of their own control could have been problematic. However, the flexibility offered by hybrid cloud models has negated that one disadvantage. Since many companies think of the Cloud as their mainframe just running somewhere else, network latency is obviously something that has to be considered to ensure the proper performance is obtained. It is thus important to work
with organizations who understand those parameters and characteristics of cloud deployment as you consider the cloud as your landing platform.
As we can see, there are many options to consider along the mainframe migration continuum. As stated earlier, there is no one perfect landing spot for every organization. It is imperative that an organization considering some sort of migration effort to think about the advantages and disadvantages of the various alternatives and see what positions them for their best possible outcome. Here at Astadia, we have helped over 200 companies find their right landing spot on the migration continuum and understand that it is one of the most important choices you will make as you consider mainframe migration as an alternative.