3.4 Resiliency Flashcards
High availability
Redundancy doesn’t always mean always available
– May need to be powered on manually
* HA (high availability)
– Always on, always available
* May include many different components
working together
– Active/Active can provide scalability advantages
* Higher availability almost always means higher costs
– There’s always another contingency you could add
– Upgraded power, high-quality server components, etc.
Server clustering
Combine two or more servers
– Appears and operates as a single large server
– Users only see one device
* Easily increase capacity and availability
– Add more servers to the cluster
* Usually configured in the operating system
– All devices in the cluster commonly use the same OS
Load balancing
Load is distributed across multiple servers
– The servers are often unaware of each other
* Distribute the load across multiple devices
– Can be different operating systems
* The load balancer adds or removes devices
– Add a server to increase capacity
– Remove any servers not responding
Site resiliency
Recovery site is prepped
– Data is synchronized
* A disaster is called
– Business processes failover to the alternate
processing site
* Problem is addressed
– This can take hours, weeks, or longer
* Revert back to the primary location
– The process must be documented for
both directions
Hot site
An exact replica
– Duplicate everything
* Stocked with hardware
– Constantly updated
– You buy two of everything
* Applications and software are constantly updated
– Automated replication
* Flip a switch and everything moves
– This may be quite a few switches
Cold site
No hardware
– Empty building
* No data
– Bring it with you
* No people
– Bus in your team
Warm site
Somewhere between cold and hot
– Just enough to get going
* Big room with rack space
– You bring the hardware
* Hardware is ready and waiting
– You bring the software and data
– Geographic dispersion
* These sites should be physically different than the
organization’s primary location
– Many disruptions can affect a large area
– Hurricane, tornado, floods, etc.
* Can be a logistical challenge
– Transporting equipment
– Getting employee’s on-site
– Getting back to the main office
Platform diversity
Every operating system contains potential security issues
– You can’t avoid them
* Many security vulnerabilities are specific to a single OS
– Windows vulnerabilities don’t commonly affect Linux
or macOS
– And vice versa
* Use many different platforms
– Different applications, clients, and OSes
– Spread the risk around
Multi-cloud systems
There are many cloud providers
– Amazon Web Service, Microsoft Azure,
Google Cloud, etc.
* Plan for cloud outages
– These can sometimes happen
* Data is both geographically dispersed and
cloud service dispersed
– A breach with one provider would not affect the
others
– Plan for every contingency
Continuity of operations planning (COOP)
- Not everything goes according to plan
– Disasters can cause a disruption to the norm - We rely on our computer systems
– Technology is pervasive - There needs to be an alternative
– Manual transactions
– Paper receipts
– Phone calls for transaction approvals - These must be documented and tested before a
problem occurs
Capacity planning
Match supply to the demand
– This isn’t always an obvious equation
* Too much demand
– Application slowdowns and outages
* Too much supply
– You’re paying too much
* Requires a balanced approach
– Add the right amount of people
– Apply appropriate technology
– Build the best infrastructure
People
Some services require human intervention
– Call center support lines
– Technology services
* Too few employees
– Recruit new staff
– It may be time consuming to add more staff
* Too many employees
– Redeploy to other parts of the organization
– Downsize
Technology
Pick a technology that can scale
– Not all services can easily grow and shrink
* Web services
– Distribute the load across multiple web services
* Database services
– Cluster multiple SQL servers
– Split the database to increase capacity
* Cloud services
– Services on demand
– Seemingly unlimited resources (if you pay the money)
Infrastructure
The underlying framework
– Application servers, network services, etc.
– CPU, network, storage
* Physical devices
– Purchase, configure, and install
* Cloud-based devices
– Easier to deploy
– Useful for unexpected capacity changes
Recovery testing
Test yourselves before an actual event
– Scheduled update sessions (annual, semi-annual, etc.)
* Use well-defined rules of engagement
– Do not touch the production systems
* Very specific scenario
– Limited time to run the event
* Evaluate response
– Document and discuss
