Friday, February 3, 2023
HomeSoftware DevelopmentVital Key Ideas and Terminologies – Be taught System Design

Vital Key Ideas and Terminologies – Be taught System Design

System Design is the core idea behind the design of any distributed methods. System Design is outlined as a course of of making an structure for various parts, interfaces, and modules of the system and offering corresponding knowledge useful in implementing such parts in methods.

On this article, we’ll cowl the usual phrases and key ideas of system design and efficiency, reminiscent of:

  • Latency, 
  • Throughput, 
  • Availability,
  • Redundancy,
  • Time
  • CAP Theorem
  • Lamport’s Logical Clock Theorem.
Important Key Concepts and Terminologies In System Design – Learn System Design

Vital Key Ideas and Terminologies In System Design – Be taught System Design

Allow us to see them one after the other.

Throughput in System Design

Throughput is outlined because the measure of quantity of information transmitted efficiently in a system, in a sure period of time. In easy phrases, throughput is taken into account as how a lot knowledge is transmitted efficiently over a time frame.



The unit of measure for throughput is bits per second or bps.

Latency in System Design

Latency is outlined because the period of time required for a single of information to be delivered efficiently. Latency is measured in milliseconds (ms).



All of us have encountered conditions the place web sites and net purposes take longer to reply, or there’s buffering whereas enjoying a video regardless of having good community connectivity. Then the latency for such methods is claimed to be comparatively excessive.

There’s a sure period of time required for person enter over the web site and there’s a sure period of time for the response from the online software to the person. So the delay between person enter and net software response to the identical enter is called latency

Causes for top Latency

Now you have to be questioning in regards to the components which can be answerable for delays. So excessive latency primarily will depend on 2 components:

  1. Community Delays 
  2. Mathematical Calculation Course of Delays 
Network Latency

Community Latency

In monolithic structure, as we all know there’s solely a single block and all community calls are native inside therefore community delay is zero and therefore latency equals computational delays solely (which if not latency equals zero in monolithic methods)

Latency = Mathematic Calculation Delays

In distributed methods, there’s a networks over which indicators are handed backward and forward therefore there’ll for positive be community delay.

Latency = Mathematic Calculation Delays + Community Delays 

Allow us to lastly talk about are parts affecting latency:

Parts Affecting Latency:

  1. Packet Dimension: Smaller the packet chunk dimension quicker the transmission and the decrease the latency.  
  2. Packet Loss: Transmission of giant packets of assorted sizes in medium losses to only a few losses in packets.  
  3. Medium of transmission: Optical fiber is the quickest means of transmission. 
  4. Distance between Nodes: Poor sign will enhance latency and nice connectivity decreases to a better extent.
  5. Sign energy: Good sign energy reduces latency.
  6. Storage delays: Saved info in a database and fetching from it requires little or no time which helps growing latency.
Causes of high network Latency

Causes of excessive community Latency

How one can Cut back latency:

  1. Use a content material supply community (CDN): CDNs assist to chop down on latency. With a purpose to shorten the gap between customers and scale back the period of time that knowledge should journey over nice distances, CDN servers are located at varied places.
  2. Upgrading pc {hardware}/software program: Bettering or fine-tuning mechanical, software program, or {hardware} parts might help lower down on computational lag, which in flip helps lower down on latency.
  3. Cache: A cache is a high-speed knowledge storage layer utilized in computer systems that briefly retailer massive quantities of transient knowledge. By caching this knowledge, subsequent requests for it may be fulfilled extra shortly than if the info had been requested immediately from its authentic storage location. This lessens latency as nicely.

Functions of Latency

  1. Autos
  2. Capital Market

Availability in System Design

Availability is the share of time the system is up and dealing for the wants. 

It’s a crucial issue when for tech corporations to offer companies whereas designing methods. As recorded, Meta went down for six hours, akin to lack of estimated 60 million {dollars}.

There are ranges related to availability with respect to the service the system is providing. As an illustration, air site visitors management requires the next degree of availability compared to the restaurant reservation system.

How availability is measured?

How availability is measured?

How availability is measured?

Now you have to be fascinated with what are these ranges and the way they’re measured. Ranges in availability are measured through downtime per yr through order of ‘nines’. Extra ‘nines’ result in lesser downtime. 

It’s as proven under through desk as follows:

Availability(%) Downtime/12 months
90 ~36.5 days 
99 ~3.65 days 
99.9 ~8.7 Hours 
99.99 ~52 Minutes
99.999 ~6 Minutes

Notice: 5 nines is taken into account because the golden commonplace of availability the system is accessible(up) to carry out duties.

How one can enhance Availability?

  1. Remove SPOF(main and vital)
  2. Confirm Automated Failover
  3. Use Geographic Redundancy
  4. Proceed upgrading and bettering

From the above understanding, we are able to wind up with two conclusions:

  1. Availability is low in monolithic structure on account of SPOF.
  2. Availability is excessive in distributed structure on account of redundancy.

Redundancy in System Design

Redundancy is outlined as an idea the place sure entities are duplicated with goal to scale up the system and scale back over all down-time

For instance, as seen within the picture under, we’re duplicating the server. So if one server goes down, then we have now a redundant server in our system to stability the load.

Redundancy in System Design

Redundancy in System Design

Now from the above picture, you have to be questioning how these connections are dealt with that means how one can get all load over to a different and never let to connect with server/s which can be already down. Right here we introduce a brand new time period referred to as the load balancer

Load Balancer

A load balancer works as a “site visitors cop” sitting in entrance of your server and routing shopper requests throughout all servers. It merely distributes the set of requested operations (database write requests, cache queries) successfully throughout a number of servers and ensures that no single server bears too many requests that result in degrading the general efficiency of the applying. A load balancer generally is a bodily gadget or a virtualized occasion working on specialised {hardware} or a software program course of. 
Take into account a situation the place an software is working on a single server and the shopper connects to that server immediately with out load balancing. It is going to look one thing just like the one under.

Load Balancer

Load Balancer

How one can deal with the unavailability of a Load Balancer?

If the load balancer turns into unavailable, then the corresponding server will turn into unavailable and the system will go into downtime. With a purpose to deal with such instances, we do go for two methods:

  1. Approach 1: Utilizing backup load balancer method: It accommodates main and secondary load balancers involving ideas of ‘floating ID’ and ‘Well being examine’.  
  2. Approach 2: Utilizing DNS Server: For now newbies to grasp affiliate this works fairly just like the redundancy precept.

Notice: Geeks do do not forget that DNS doesn’t hold monitor of  whether or not load balancer is working so will we introduce Monitor  to maintain monitor in case of DNS. 

Tip: In system designing failures are inevitable, we couldn’t eliminite them fully can solely  work in minimizing them. 

Whereas contemplating the provision issue whereas designing methods is immediately proportional to geographic location. As an illustration: For a corresponding service if a system goes down in a selected location (be India) the entire service is accessible at another location to make the service operational. In the actual world, we’re having the whole {hardware} out there throughout varied places in order to not hamper service at any price.    

Consistency in System Design

Consistency is known as knowledge uniformity in methods

When a person requests knowledge, the system all the time returns the identical knowledge, whatever the person’s location, time, and many others. Earlier than a person receives knowledge from any node, the server at which the info is up to date or modified ought to efficiently replicate the brand new knowledge to all of the nodes.                         

With a purpose to perceive consistency in less complicated phrases:

  • Take into account three nodes X, Y, and Z in a system, the place little or no t1 time is taken for knowledge transmission for replication from X to Y and t2 from X to Z
  • Now the utmost amongst t1 and t2 is taken the place no person is meant to fetch data in order to keep away from to get inconsistency in knowledge by offering older information.    

Instance: Account Transactions

Notice:  If the customers will get the info in some way beccause of inconsistency in methods, it is called Soiled Studying.                            

Time in System Design

Time is a measure of sequences of occasions occurring which is measured right here in seconds in its SI unit. 

It’s measured utilizing a clock which is of two varieties:

  1. Bodily Clock: answerable for the time between methods.  
  2. Logical Clock: answerable for the time inside a system.  
Time in System Design

Time in System Design


Listing out all schools on this planet. So right here the time taken between schools for communication is a measure of the bodily clock. Whereas the time taken inside the hospital is a measure of a logical clock.  

Three fascinating traits of distributed methods with replicated knowledge are known as CAP: partition tolerance, availability, and consistency (amongst replicated copies) (within the face of the nodes within the system being partitioned by a community fault). In line with this theorem, in a distributed system with knowledge replication, it isn’t potential to make sure all three of the required properties—consistency, availability, and partition tolerance—on the identical time. It claims that solely two of the three properties said under may be supported strongly by networked shared-data methods:

  • Consistency
  • Availability
  • Partition Tolerance  
CAP Theorem

CAP Theorem

The completely different mixtures and their use instances are as follows:

  1. CA: Consistency and Availability:
  2. AP: Availability and Partition Tolerance: 
  3. CA: Consistency and Availability:  
CAP Theorem

CAP Theorem

Lamport’s Logical Clock is a course of to establish the sequence during which occasions happen. It acts as the muse for the extra complicated Vector Clock Algorithm. A logical clock is required as a result of a distributed working system (Lamport) lacks a worldwide clock.

Bear in mind: Leslie Lamport got here up with the concept for Lamport’s Logical Clock. 

Tip: Refer right here to learn extra and perceive higher the Lamport Logical Clock Theorem.  



Please enter your comment!
Please enter your name here

eighteen − thirteen =

Most Popular

Recent Comments