TOPIC : GOOGLE CLOUD PLATFORM

• Subject Name : Cloud Computing

• Subject Code : PEC-CS701B
• Continues Assessment 1 (CA1)
• Student Name : Apratim Mukherjee
• Stream : Computer Science & Engineering
• University Roll : 17600120004
• 7th Sem, 4th Year
 2

 WHAT IS CLOUD COMPUTING ?

In Simplest terms, cloud computing means storing and accessing the data and programs on remote servers that are
hosted on the internet instead of the computer’s hard drive or local server. Cloud computing is also referred to as
Internet-based computing.
Cloud Computing Architecture: Cloud computing architecture refers to the components and sub-components
required for cloud computing. These components typically refer to:
o Front end(fat client, thin client)
o Back-end platforms(servers, storage)
o Cloud-based delivery and a network(Internet, Intranet, Intercloud).
Hosting a cloud: There are three layers in cloud computing. Companies
use these layers based on the service they provide.
 Infrastructure
 Platform
 Application

Apratim Mukherjee || University Roll : 17600120004

 3

 WHAT IS GOOGLE CLOUD?

Starting from 1998 with the launch of Google Search. google has developed one of the largest and most
powerful IT Infrastructures in the world. Today, this infrastructure is used by billion of users to use services
such as Gmail, Youtube, Google Photo, and Maps. In 2008, Google decided to open its network and IT
infrastructure to business customers, taking an infrastructure that was initially developed for consumers’
application to public service and launching the google cloud platform.

Google provides all the services listed above, hence the name Google Cloud Platform (GCP). Apart from these,
there are so many other services provided by GCP and many concepts related to it that we will discuss in this
article.

 REGIONS, ZONES, RESOURCES, QUOTA :

All generally available Cloud Platform resources that have regional scope, such as Compute Engine,
are available in all regions/zones. For products that have global scope, such as App Engine and BigQuery,
you do not need to select a region or zone. Let’s list the regions where Compute Engine (and, by extension,
persistent disks, load balancers, autoscalers, Cloud Storage, Cloud Datastore, and Cloud SQL) is available,
using gcloud:

Apratim Mukherjee || University Roll : 17600120004

 4
Cloud compute regions list
NA$ME CPUS DISKS_GB ADDRESSES RESERVED_ADDRESSES STATUS TURNDOWN_DATE
• asia-east1 2.00/24.00 10/10240 1/23 1/7 UP
• europe-west1 0.00/24.00 0/10240 0/23 0/7 UP
• us-central1 0.00/24.00 0/10240 0/23 0/7 UP

$ Cloud compute zones list

• NAME REGION STATUS NEXT_MAINTENANCE TURNDOWN_DATE
• asia-east1-a asia-east1 UP
• asia-east1-c asia-east1 UP
• asia-east1-b asia-east1 UP
• europe-west1-b europe-west1 UP
• europe-west1-c europe-west1 UP
• europe-west1-d europe-west1 UP
• us-central1-f us-central1 UP
• us-central1-a us-central1 UP
• us-central1-c us-central1 UP
• us-central1-b us-central1 UP
This output shows that there are a total of 10 zones across 3 regions. Of course, this is as of this writing;
Google is expected to add new regions and zones regularly.
From the region and zone names, you can decipher that the fully qualified name for a zone is made up of
<region>-<zone>. For example, the fully qualified name for zone a in region us-central1 is us-central1-a.

Apratim Mukherjee || University Roll : 17600120004

 5

 APIS AND CLOUD CLIENT LIBRARIES :

Some people are afraid to bring their workloads to the cloud because they’re afraid they’ll get locked into a specific
vendor. But in many ways, Google gives customers the power to run their applications elsewhere, if Google
becomes not the simplest provider for his or her needs. Here are some samples of how Google helps its customers
avoid feeling locked in. GCP services are compatible with open source products. For example, take Cloud Bigtable,
a database that uses the interface of the open-source database Apache HBase, which provides customers the
advantage of code portability. Another example, Cloud Dataproc provides the open-source big data environment
Hadoop, as a managed service, etc.
Google follows an API-first development philosophy, and APIs are the primary
developer interface for
Google’s products, including Cloud Platform. Hence, before you can use a product
—say, Compute
Engine—you need to enable that particular API in your project. API enablement is
on a project-by-project
basis. Google makes it easy for you to enable a particular API using the Developers
Console. You can access
the APIs section by choosing Developers Console ➤ APIs & Auth ➤ APIs. The
tabbed screen shows the list
of all available APIs and the APIs that have been enabled in a project.

Apratim Mukherjee || University Roll : 17600120004

 6

 CLOUD PLATFORM PRODUCTS:

• Compute
o Compute Engine: Compute Engine is an infrastructure-as-a-service (IaaS) product. Using it, you can launch virtual machines, create networks, and attach local
and remote persistent disks based on magnetic or solid-state technologies.
You can also design and build advanced architectures that include load balancing and auto-scaling and that span multiple zones in a region or multiple geographical
regions worldwide. Compute Engine gives you maximum
flexibility and is primarily targeted at architects and system administrators.
o App Engine: App Engine is a platform as a service (PaaS) product. Using it, you can build web-scale, autoscaling applications. App Engine is targeted at
software developers and provides a comprehensive collection of libraries. Using it, you can simply upload an application to the platform, and App Engine takes
care of everything else.
o Container Engine: Containerized applications are being explored as the next

step in DevOps standard operating procedures and the next generation of

application development. Docker is at the forefront of this revolution and
is building an industry-wide consensus about the format and interface of
application containers. An application container is enabled by a set of core
innovations in the Linux kernel that Google invented almost a decade ago. This
places Google at the forefront of driving container adoption among developers.
Container Engine is covered in Chapter 6; it is still in the early stage of evolution.
Apratim Mukherjee || University Roll : 17600120004
 7

o Managed VMs: Managed virtual machines are the next generation of App Engine and feature many new capabilities such as Docker-formatted
application containers, writable local disks, and live debugging of applications over SSH. Whereas Container Engine enables you to build
sophisticated multi-tier applications where each node is a Docker container, managed VMs take care of all of them.

• Storage
o Cloud SQL: Cloud SQL is a managed RDBMS product and is 100% binary compatible with open-source MySQL server software. Google
manages all the database-management tasks, and you can focus on building an app that needs a SQL back end. Cloud SQL supports advanced
configurations such as read replicas (internal and external) and SSL connections.
o Cloud storage: Cloud storage is object-based file storage that you can use to
store data files without worrying about file system setup and maintenance.
Cloud storage also includes automatic transparent global edge caching so that
you don’t have to set up another entity manually. Cloud storage offers different
product flavors based on durability characteristics.
o Cloud Datastore: Cloud Datastore is a managed, NoSQL, schemaless database
for storing non-relational data. You can use this service to store key:value-based
data. Cloud Datastore scales as your data needs increase, and you pay only for the
space that you consume.

Apratim Mukherjee || University Roll : 17600120004

 8
• Big Data
o BigQuery: BigQuery is a hosted Big Data analytics platform.
BigQuery lets you query datasets that are multiple terabytes in
size and features data ingestion at the rate of 100,000 rows per
second per table.
o Cloud Pub/Sub: Cloud Pub/Sub is a hosted messaging
and queuing products that lets you connect multiple producers
and consumers and enable low latency, high-frequency
data transfer between them.
o Cloud Dataflow: Cloud Dataflow is a simple, flexible,
powerful system you can use to perform data-processing
tasks of any size. It lets you build, deploy, and run complex
data-processing pipelines.

• Services
o Cloud Endpoints: Cloud Endpoints enables you to create RESTful
services and make them accessible to iOS, Android, and JavaScript
clients. It also automatically generates client libraries to make
wiring up the front end easy. With built-in features include denial-
of-service protection, OAuth 2.0 support, and client key
management, Cloud Endpoints lets you host API endpoints in
Cloud Platform.
o Google APIs: Applications can consume both Cloud Platform
product APIs (for example Google Storage) and Google products
APIs (for example Google Maps). This book includes an example
of using the Translate API to translate content among 90 pairs of
human languages
Apratim Mukherjee || University Roll : 17600120004
 9

• Networking

o Cloud DNS: Cloud DNS is a reliable, resilient, low-latency DNS service from Google’s
worldwide network of Anycast DNS servers. You can manage your DNS records using the
Developers Console UI, the gcloud command-line tool, or a full-featured RESTful API.
o •Authentication: Authentication is an essential step for governing access to your Cloud
Platform resources or Google user data. Google uses the OAUTH 2.0 protocol exclusively
for both authentication and authorization.

 PRICING :

Google was the primary major Cloud provider to bill by the second instead of rounding
up to greater units of your time for its virtual machines as a service offering. This may not
sound like a big deal, but charges for rounding up can really add up for customers who
are creating and running lots of virtual machines. Per second billing is obtainable for a
virtual machine use through Compute Engine and for several other services too.

Compute Engine provides automatically applied use discounts which are discounts that
you simply get for running a virtual machine for a big portion of the billing month. When
you run an instance for at least 25% of a month, Compute Engine automatically gives
you a reduction for each incremental minute you employ it. Here’s one more way
Compute Engine saves you money.
Apratim Mukherjee || University Roll : 17600120004
 10
 ADVANTAGES :
1. Good documentation: We are talking about many pages in total, including a reasonably detailed API Reference guide.
2. Different storage classes for every necessity: Regional (frequent use), Nearline (infrequent use), and Coldline (long-term
storage).
3. High durability: This suggests that data survives even within the event of the simultaneous loss of two disks.
4. Many regions available to store your data: North Ameria, South America, Europe, Asia, and Australia.
5. The “Console” tab within the documentation allows you to try for free of charge different SDKs. It’s incredibly useful for
developers
6. One of the simplest free layers within the industry. $300 free credit to start with any GCP product during the primary year.
Afterward, 5 GB of Storage to use forever without any charges.
 DISADVANTAGES:
1. The support fee is sort of hefty: Around 150 USD per month for the foremost basic service (Silver class).
2. Downloading data from Google Cloud Storage is expensive. 0, 12 USD per GB.
3. Google Cloud Platform web interface is somewhat confusing. Sometimes I am lost while browsing around the menus.
4. Prices in both Microsoft Azure (around 0.018 USD per GB/month) or Backblaze B2 (about 0.005 USD per GB/month) are less than Google
Cloud Storage.
5. It has a high pricing schema, almost like AWS S3, so it’s easy to urge unexpected costs (e.g. number of requests, transfers, etc.).

Apratim Mukherjee || University Roll : 17600120004

 11

Apratim Mukherjee || University Roll : 17600120004