Reference architectural decisions to migrate IBM i on-premises to IBM Power Systems Virtual Server
This chapter provides an example of migrating IBM i workloads from on-premises to off-premises IBM Power Systems Virtual Server on IBM Cloud.
We describe the considerations, architectural decisions, issues or problem statements, assumptions, motivations, alternatives, and implications that are part of such a process.
This chapter includes the following topics:
7.1 Introduction
In this section, we describe important architectural considerations before migrating IBM i on-premises workloads to off-premises. We discuss one of the most common methods that is used to migrate IBM i environments to the cloud.
We also include a use case that is based on real customers experiences, which includes functional and non-functional requirements. Then, an overview of an architecture diagram is presented.
 
Note: For more information about customer’s current IBM i systems, see this Seismic web page (log in required).
7.2 Use case overview
Suppose that a customer signed an agreement with IBM to move 10 IBM i virtual machines (VMs) that are in La Paz - Bolivia and named LPZ01 and LPZ02 to IBM Power Systems Virtual Server on IBM Cloud in SAO01 and SAO04, both in Sao Paulo, Brazil.
Per the agreement signed with the customer, all IBM i VMs are to be moved “as is”, with operating system release 7.1. No transformation is included in the project, except the operating system upgrade to 7.3.
The following VMs are duplicated to SAO04:
Four production IBM i VMs
One development VM
One archive VM
Four Disaster Recovery (DR) VMs
This environment includes storage, backups in IBM Cloud Object Storage, and an internal network, which features firewalls and a jump server for the IBM i VMs access management.
Also, thus use case includes a third-party vendor logical replication solution between the source and target, which replicates across different zones in IBM Cloud (see Figure 7-1).
Figure 7-1 Migrating IBM i VMs from La Paz (on-premises) to Sao Paulo (off-premises)
 
Note: Bolivia and Brazil are in the same continent (South America); therefore, this use case includes multiple zones. The cities of La Paz and Sao Paulo cities are separated by 3000 kilometers (1684 miles). For more information about multi-zone regions, see this LIBM Cloud Docs web page.
7.2.1 Scope
In this use case, all 10 IBM i VMs must be migrated from two data centers (LPZ01, which is the source, and LPZ02, which is the target) and are managed by a customer in La Paz, Bolivia, to the IBM Cloud data center in SAO01 (the source) and SAO04 (the target).
Development and archive VMs that are in the La Paz-Bolivia data center are to be migrated to IBM Cloud by using IBM Cloud Object Storage backup. Then, they are to be restored to the target VM in IBM PowerVS on IBM Cloud. The IBM i production VMs that are on LPZ01 are to be migrated by using the IBM Cloud Object Storage backup and logical replication method.
 
Note: A maintenance contract, an activation key of the logical replication, and installation and configuration that done by the third-party vendor often are needed for this process.
After the IBM i VMs are active, the data is restored to the target location and the connectivity from La Paz data center is established by using IBM i Access Client Solution (ACS). After connectivity is validated, the customer’s applications and their functions are configured in the environment.
Also, each VM and the overall function for each environment are tested.
 
IBM provides to customer the temporary technical information. such as: IP addresses, DNS names, and VM temporary name. This information required so that the tests can be completed without any errors on IBM i production environments that are established in La Paz Bolivia.
The customer provides the firewall rules that are to be set up by IBM. User profiles are automatically transferred to the IBM Power Systems Virtual Server on IBM Cloud that is restoring user profiles (RSTUSRPRF). It is the customer’s responsibility to administer their applications and user ID management.
After migration, IBM can run the suitable security scan for each IBM i environment that is based on the customer’s security policy. IBM also provides the evaluation results to the customer.
Often, the following teams (among others) are involved in the transition and steady state support of the environment:
IBM Cloud connect for the network
IBM Cloud Object Storage
IBM PowerVS
IBM i support team
Customer’s application team
Third-party vendor software team support for logical replication
Image catalogs are created out of objects that are backed up by using optical devices. These catalogs must be restored on the IBM Power Systems Virtual Server instance by using some of the migration strategies that us IBM Cloud Object Storage and NFS server.
IBM Backup, Recovery, and Media Services (BRMS) is an IBM i product that can be used to automate tasks that help define and process your backup, recovery, and media management operations. IBM Content Collector can be integrated with IBM BRMS to move and retrieve objects from remote locations, including IBM Cloud Object Storage.
The following procedure shows how the customer’s IBM i operating system and data is migrated from an on-premises system to the IBM Cloud environment. (Most of these steps can be automated by using IBM BRMS and IBM Content Collector):
1. Install IBM BRMS and IBM Content Collector.
2. Install IBM i System Minimum PTF levels.
3. Install BRMS and IBM Cloud Object Storage PTFs.
4. Install Cloud-Init for IBM i and PTFs.
5. Save the system by using IBM BRMS and IBM Content Collector to IBM Cloud Object Storage.
6. Create IBM BRMS Recovery Reports.
7. Copy the data to IBM Cloud Object Storage.
8. Copy the data from IBM Cloud Object Storage.
9. Build a VM in PowerVS by using NFS VM.
10. . Restore the user data by using the image catalog.
Infrastructure considerations
The technical infrastructure architecture and design consist of deploying several zones in IBM Cloud in Sao Paulo. In addition to the IBM Power Systems Virtual Servers zone that is named Power Colo (Power Collocation), a front-end zone is created in which jump servers are deployed for users access and other functions, such as a proxy for accessing the IBM Cloud Object Storage services. These services are hosted in a bare metal server.
A cluster of firewalls is deployed in the front-end zone within the SAO01 site. A stand-alone firewall is deployed in SAO04.
Functional requirements
The solution must satisfy functional and nonfunctional requirements in a way that best balances competing stakeholders’ concerns and that considers any relevant constraints (see Table 7-1).
Table 7-1 Functional requirements
Requirement
Description
Management services for IBM i.
Support for IBM i operating systems.
Provide backup services.
IBM Cloud Object Storage is used as backup-services.
Provide multi-site HA solution.
Multi-site infrastructure to be provided in Sao Paulo (dual sites).
Dual sites high availability.
The use of any third-party vendor solution on logical replication to establishing DR between SAO01 and SAO04.
Provide fault tolerant LAN infrastructure in IBM Cloud.
Provide network connectivity for application and servers.
Data replication between IBM Cloud and customer data center.
The customer uses a logical replication solution to replicate the data for the IBM i application.
Provide traffic isolation and segmentation.
The use of jump servers and traffic filtering on IBM Cloud.
Provide WAN connectivity.
Customer provides WAN circuit and the POP network infrastructure; IBM provides the termination endpoint in Sao Paulo.
Nonfunctional requirements
The following nonfunctional requirements must be met:
IBM Cloud portal access for IBM i VMs provisioning.
Worldwide Tools Solutions for alert monitoring and reporting (IBM i).
Traffic bandwidth in IBM Cloud infrastructure does not exceed 1 Gbps.
Traffic bandwidth for replication is limited to 500 Mbps. Internet is used for preserving production traffic.
Local network redundancy to be provided in primary IBM Cloud site (SAO01); firewall cluster in High Availability, dual ports connectivity.
Manageability access for customer’s users to be provided by using jump servers.
 
Note: For more information about Worldwide Tools Solutions (WWTS), see this IBM Support web page.
7.2.2 Architectural decisions
Architectural decisions help to communicate why the solution architecture is used across the scope of the solution. More than one conceivable arrangement choice is available for a specific architectural issue.
This choice incorporates components and their connections, innovation options, allotting usefulness to different components, making situation choices for components that are facilitated inside different infrastructure nodes, and so on.
The choices can have diverse costs that are related to them, the degree they fulfill different prerequisites. and can show distinctive ways of adjusting competing stakeholders’ concerns. Architects can perform the following tasks:
Formally archive the basic choices that they make in creating the arrangement
Agree as to why the arrangement looks the way it does.
In summary, architectural decisions focus on the following factors:
Explaining the method that was chosen and why that method is being used.
Offer assistance to ensure that the arrangement fulfills functional and nonfunctional prerequisites. If ti does not fulfill those requirements, offer assistance to make it unequivocal to the partners.
Avoid superfluous adjustments through the arrangement conveyance lifecycle.
The following tables list some example decisions about lifting and shifting IBM i VMs from the various architectural perspective:
Table 7-2: Infrastructure
Table 7-2 Infrastructure: Front-end accounts
Architectural decision
Front-end accounts that are used for accessibility and provisioning some services.
Problem statement
Providing a way for accessing the target Power VS, which is moved from the customer’s data centers to IBM Cloud in Sao Paulo in a dual site configuration.
Assumptions
Customer provides the WAN connectivity up to the network PoP Equinix1 that is next to the data center.
Motivation
Standard design for this type of solution.
Alternatives
None.
Decision
Deploy front-end account and services.
Justifications
For accessibility of the PowerVS, a front-end zone is needed.
Some services are provided, such as the control of network traffic (firewalls), a relay environment for accessing the target IBM i images, and a Proxy for IBM Cloud Object Storage access.
Implications
Deploy WAN access and replication method for moving the data in the target environment.
Derived requirements
Provide firewall services for VPN access and filtering of traffic.
Provide IBM Cloud Object Storage services for Backup.
Provide WAN network connectivity for customer’s users and application connectivity.
Provide bare metal servers to hosts relay applications and proxy.
1For more information about Equinix, see America Data Centers.
 
Important: Consider the following points:
By using Equinix, you can get a direct link to reach IBM Power Systems Virtual Server over Direct Link Connect. From Equinix, you also can get a cross connection to Megaport and connect to IBM Power Systems Virtual Server directly.
Before you begin, determine the location connection to IBM Cloud by verifying your colocation provider’s or service provider’s capabilities to reach the Meet-Me Room and cross-connect into IBM Cloud. For more information, see this web page.
On SAO01, the location type and Meet-Me Room (MMR) Operator, is DC and Digital Realty (Ascenty), and SAO04 DC(AZ1) and ODATA.
In this example, to lift and shift IBM i VMs from Bolivia to Brazil, the connection from Bolivia to SAO01 can be established. For example, you must contract directly with a carrier that can provide enough capacity and presence in any Ascenty data center. The solution might be LAN-to-LAN + Cross Connection Fiber Optic + IBM DirectLink 1 G or 10 G.
For the LAN-to-LAN link, IBM must directly contract with the carriers for the private LAN-to-LAN circuits.
For example, the following carriers and the location of the Ascenty data centers are used:
 – ALGAR TELECOM
 – ALOO TELECOM
 – AMERICANET
 – ANGOLA CABLES
 – BRDIGITAL, LUMEN
 – COGENT
 – SEABORN
 – CLARO - PRIMESYS - EMBRATEL
 – Durand/Tavola
 – HOSTFIBER
 – MEGATELECOM
 – MUNDIVOX
 – NETELL
 – NEUTRONA NETWORKS
 – NOVVACORE TELECOMUNICAÇÕES
 – OI MOVEL
 – Seaborn
 – SILICA NETWORKS
 – VIVO
 – VOGEL
 – WCS TELECOM
For more information about Digital Realty (Ascenty) and ODATA, see the following web pages:
 – Ascenty
 – ODATA
Table 7-3 Infrastructure: Dual site
Architectural decision
Dual site infrastructure is required for HA purposes.
Problem statement
If a major outage occurs, the customer’s users can connect to back up site (use of DNS for servers translation; secondary site has a different TCP IP address).
Assumptions
Two sites are used for the solution: one in SAO01 and the other in SAO04 (in a different zone).
Motivation
Infrastructure recovery if a major outage occurs.
Alternatives
None.
Decision
Deploy dual site solution in an IBM Cloud Multi-Zone Region (Sao Paulo).
Justifications
If a major outage occurs at the primary site, the main goal is to restart part of the application and services in the secondary site.
Implications
Deploy a secondary site in addition to the Production environment.
Derived requirements
Provide WAN network connectivity to secondary site for customer’s users and application connectivity.
Duplicate part of the primary infrastructure in backup site.
Table 7-4 Migration strategy and backup
Architectural decision
IBM Cloud Object Storage backup is used for migrating IBM i VMs to SAO01 and SAO04.
Problem statement
Back-up and data replication between client data center and IBM Cloud target infrastructure. No Automatic Tape Library (ATL) or VTS is available to perform a save and restore, which is a traditional migration method for IBM i operating system.
Assumptions
The use of IBM Cloud Object Storage for the migration is one of the available methods for moving workload to IBM Power Systems Virtual Server in IBM Cloud.
Motivation
The use of IBM Cloud Object Storage to move IBM i workloads to SAO01 and SAO04.
Alternatives
The use of IBM Cloud Object Storage for migration.
The use of Master Data Management (MDS) device for the migration.
Transferring IBM i image OVA file to IBM Cloud Storage by using IBM Power Virtualization Center.
Decision
IBM Cloud Object Storage is used.
Justifications
Master Data Management is excluded because of the delay for taking backups data on 1 Gbps interface.
Customer does not have a virtualization by PowerVC.
Implications
Network connectivity includes VPN WAN connectivity and Proxy in front-end account.
Derived requirements
Deploy Proxy in Front-End zones and VPN access from client on IBM Cloud.
Buckets are needed to create on IBM Cloud Object Storage for the data move. Needs more storage for the IBM Cloud Object Storage backup in the source IBM i VM.
Table 7-5 Delta data synchronization
Architectural decision
A third-party vendor replication solution is used.
Problem statement
Customer has IBM i VMs with storage 10 TB - 70 TB. Some of the IBM i VMs include a journal of 1 TB that generates daily. The use of a logical replication tool is the best solution to remove the delta data after the IBM i VM restoration on the target is complete.
Assumptions
The logical replication tool syncs up the data between source system and target system.
Motivation
Customer uses a logical replication.
Alternatives
PowerHA for IBM i geographic mirroring.
Decision
The use of the current third-party vendor tool.
Justifications
In this case, this option is the most suggested because customer has a third-party logical replication tool in support.
Implications
Third-party vendor provides license key for the logical replication tool to be deployed on IBM PowerVS. The third-party vendor installs the tool in the cloud.
Derived requirements
The third-party vendor provides a temporary license key to migrate data to IBM Cloud.
 
Note: PowerHA Geo mirroring in the IBM Power Systems on IBM Cloud does not include any unique features. It is host-based and done over a TCP/IP connection. The only possible unique part is that you must design the network between the two servers, but that requirement is not unique to PowerHA. Geo Mirroring includes limitations, such as it works for IASP only, with which a customer cannot be configured.
Table 7-6 Networking: IBM Cloud Direct Link Dedicated on Classic
Architectural decision
WAN direct-link connectivity to be redundant: one primary and one secondary link.
Problem statement
WAN access connectivity to be recovered in case of primary link outage.
Assumptions
WAN part is customer’s responsibility, IBM Cloud provides dual circuit connectivity on diverse physical devices.
Motivation
Maintain connectivity with customer’s corporate network.
Alternatives
Doubling the WAN connectivity: a redundant connectivity in SAO01 and a redundant connectivity in SAO04.
Decision
Provide redundant connectivity in SAO01 and use the IBM Cloud backbone for Inter-site communications.
Justifications
The provided service level is consistent and the option is available to connect the IBM Cloud site by using VPN.
Implications
Deploy Sao Paulo site-to-site connectivity.
Derived requirements
Deploy GRE and Direct link connectivity for Front-End zones communications.
 
IBM Cloud Direct Link Dedicated
IBM Cloud Direct Link Dedicated is a single-tenant product (see Table 7-6 on page 252). It offers a dedicated port that is suited for banks, insurance companies, or anyone with strict compliance policies.
A fiber cross-connection is created through a network service provider (NSP) in an IBM Cloud network Point of Presence (PoP). IBM engineers facilitate end-to-end connectivity with your selected NSP, and you can access your cloud infrastructure in the local IBM Cloud data center.
The NSP runs last-mile links directly between a router on your network and an IBM Cloud router. As with all of the Direct Link products, you can add global routing that enables private network traffic to all IBM Cloud locations.
For more information, see this IBM Technology YouTube video.
IBM Cloud Direct Link is available in the following offerings:
IBM Cloud Direct Link on Classic:
 – Direct Link Connect on Classic
 – Direct Link Dedicated on Classic
 – Direct Link Exchange on Classic
 – Direct Link Dedicate Hosting on Classic
IBM Direct Link 2.0:
 – Direct Link Connect
 – Direct Link Dedicated
For more information about which Direct Link solution to order, see the following IBM Cloud Docs web pages:
 
 
Important: Tables Table 7-2 on page 249 - Table 7-5 on page 252 list decisions only as an example. Real world decisions can vary according to the customer, scenario, third-party vendor applications, in-house applications, region, networking, and so on.
A certified IBM i architect can help make decisions about your scenario.
7.2.3 Architecture overview
Figure 7-2 shows an overview for IBM Cloud in Sao Paulo. The division of the different zones is delineated.
Figure 7-2 IBM i on IBM PowerVS architecture
These zones are separated from each other by physical switches, which is an activity that is controlled by firewalls. A dual-site infrastructure is used. Consider the following points:
SAO01 is where the IBM i production runs and it comprised in two subzones:
 – The first subzone is named Client Front End Account. Here, you can see Jump servers and services, such as IBM Cloud firewalls and proxy. The two zones communicate by using the internal IBM Cloud network backbone.
 – The second subzone is named Power Colo and it is here where IBM PowerVS is installed.
 – SAO04 is used for DR purposes if an outage occurs at SAO01.
Consider the following points:
The System Context Diagram (SCD) provides the networking environment and the IBM i VMs and jump servers. It also identifies the interfaces between entities.
An Operational Model provides a high-level description of an abstraction that specifies how the components of the application are organized and connected, where they are to be located and hosted, and what business requirements the application resolves.
The backup and recovery strategy regarding that the backups are saved in IBM Cloud Object Storage if backups are less than 2 TB to be a viable solution the use of BRMS/ICC to save to IBM Cloud Object Storage. For more information, see this IBM IPP Wiki page.
PowerHA for IBM i Geographic Mirroring or a third-party vendor for logical replication solutions can be strategies for IBM i data to be replicated across zones in the IBM Cloud.
For more information about Architectural decisions, see this IBM Training web page.
 
..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
3.137.188.11