One of the IT standard requirements would be to provide high availability to all its high priority tools. The most known terms used in this regard are – High availability (HA), redundancy, Disaster recovery (DR). While IT might have done this kind of setup for several applications but still for each application it would make sense to know “what to do, how to do”, as applications differ in many aspects. For the benefit of HCL Software customers, this post is trying to provide some useful information with respect to HCL Launch.
Phases in achieving High Availability
- Background for High availability need
- Business statement for HA/DR/redundancy to start work
- Deployment plans to represent HA
- Timelines & efforts
- Budget considerations
- Execution plans to not miss key things
Background for HA
Firstly, we should have confirmation on what we are trying to setup.
High availability (HA) –when usage is growing, performance has become important/performance issue to be solved, achieve zero downtime for business-critical applications or SLA driven applications.
Disaster Recovery (DR) – when application must be ensured recoverable against any disaster of data centres
Redundancy – when applications to be improved with response time, handle more requests, or provide expected response standards
Since each of these looks alike in some cases but key point is that we must decide what is our goal? Some organizations have mandatory compliance need to ensure DR so you must be doing that. In other case they are driven out of other requirements like growth, performance, standards etc. Once decided if you must go for High availability, we must think is it Redundancy or HA?
Deliverables: Use cases, application performance/adoption, downtime or revenue impact numbers, future growth rate metrics.
Business statement for HA
Business statement is like “we would like to start onboarding critical enterprise applications to HCL Launch” or “we would like to ensure that there is zero down time for HCL Launch”. This gives direction and clarity on why we must achieve HA. If its number of growing applications within HCL Launch which can be achieved through horizontal expansion by adding new servers or relays, then you will not require to work hard for HA. So please focus on outcomes at this point.
Deliverables: A document probably with all details of mission statement, including execution, budget, timelines, and approvals of various stake holders
Deployment plans that achieves expected business goals
Use our deployment diagrams as base reference and create one for your organization by covering the diagram with labels for server names, geographical locations, network latency between data centres, capacity etc so that the picture providers lots of information. Its good to put expected load on them by mentioning the load factors. Try to indicate information of IT asks like storage, replication need (yes/no), compute details. Label ports and communication enablement plans as well including firewalls, data center names, geographical location. [ you may need to work with network teams].
Deliverables: Representation of present and future state of application deployment with great insights covered, checklist(s) specific to your organization, checklist for future upgrades, support and other IT regular efforts.
Timelines and efforts
Create timeline plans that include procurement to deployment. Also align as per downtimes and other teams that must work along with your group. Its good practice to get timelines reviewed with all stake holders [ ensure in checklist stake holders are listed] like network, database, storage, compute, HCL Launch support.
Deliverables: timeline information, stake holder list and approvals, events and leads information.
While planning for budget consider following things
- Budget for license cost of HCL Launch [ each server would cost]
- Cost of compute, storage, database extra license
- Backup strategy and cost if application within organization as cross charge
- Man power costs
Execution plan should have following things:
- Bill of materials
- Status of support tickets, status, approvals
- Downtime window
- Roll back plan
- Checklist for testing
- Confirm the level of support you expect from HCL Launch
- Creating a support ticket to communicate upgrade plans to HCL Launch team, it helps HCL Software to provide best support
- List of Software versions selected [be aware of end of support information for versions selected]
- If you are using any other product integrations with HCL Launch, checklist to ensure their work with new deployment plan
- Checklist to enable logging at network, application, integration, plug-in etc levels to troubleshoot any issues during execution