What is IaaS?
Infrastructure as a Service frees businesses from the legacy IT infrastructure limitations, which require them to heavily invest in hardware and software. Purchasing and maintaining hardware and software is expensive, especially for small and medium businesses that don’t have the kind of IT budget large businesses and enterprises have at their disposal.
IaaS is the answer for businesses that want instant availability of virtual servers, computing power and storage capacity for their sites, marketing campaigns and so on. The on-demand, scalable infrastructure allows them to decrease or increase capacity within just a few minutes instead of waiting for days or even weeks. We have already covered benefits and challenges of SaaS in detail, which also apply to IaaS as well as other cloud deployment and service models.
From mission-critical systems to CRMs and social media apps, enterprises rely on smooth operation of these applications to stay competitive. Scalability and flexibility is often the main concern when using on-premises solutions. IaaS solutions allow enterprises to effectively and efficiently manage their workloads and offer some major benefits including:
On-demand Scalability and elasticity
IaaS providers have a lot more available capacity than most enterprises. This results in an optimized cloud resource utilization and enables businesses to scale as and when needed.
Since enterprises don’t have to set up the infrastructure themselves, they can save a lot in upfront costs and time associated with maintaining hardware/software/network and securing the premises. IaaS is also a great solution for businesses that want to test new ideas or just want to manage seasonal spikes in resource requirements.
Enterprises need low latency, high IOPS and throughput for smooth operation, which at a large scale can be very expensive if done in-house. IaaS providers already have the infrastructure needed for ensuring high performance computing, saving businesses time and money.
Reliability and Redundancy
IaaS providers have their data centers spread across the globe so businesses get protection against single point-of-failures. Pooled resources also make it possible for enterprises to carry on with their business during maintenance sessions or downtime.
Massive scalability allows enterprises to market out faster and reach the target audience in less time than their competition. This also helps businesses to focus more on growth than spending a lot of time on IT matters.
Disadvantages of IaaS
Although IaaS offers huge advantages and a competitive edge, like every other technology it has its own drawbacks, which are similar to what we have already covered here and including:
Enterprises depend on IaaS providers to keep everything up and running. If a disaster strikes, there isn’t much they can do to take complete control. An SLA defines what businesses should expect from the provider so it’s important to consider the terms and conditions before committing to anything.
Vendor lock-in is another important aspect to consider. Although proprietary solutions have their own benefits, open source/platform solutions allow enterprises to move their data easily without getting locked-in with a particular vendor.
IaaS providers go to every length to ensure security of their infrastructure and services. However, businesses still have very little control over cloud security and most of the decisions are made by the provider.
Top 5 IaaS Providers
1. Amazon Elastic Compute 2
Amazon EC2 is among the most popular IaaS service providers that offers highly scalable in-the-cloud computing services for developers. The easy-to-use interface is designed to make the capacity configuration process smooth and provides developers with complete control over the computing resources.
Based on publically published data and downtime for the year 2018, Amazon EC2 experienced 7 times fewer downtime hours than the next biggest provider. The provider caters to needs of businesses of almost all sizes and has 69 global availability zones in 22 different regions. The AWS instance categories include:
- General-purpose for business-critical apps, small and medium databases and web apps
- Compute Optimized for high-performance computing, video encoding and batch processing
- Memory Optimized for high performance DBs, big-data/real-time analytics, distributed web
- Accelerated Computing for graphic intensive apps, machine learning and gaming
- Storage Optimized for NoSQL DBs, distributed file sys and data warehousing
EC2 is free-to-try and the free tier includes 750 hours of Windows and Linux t2.micro instances per month (for a year, users have to use EC2 Micro Instances to stay within the tier). Flexible pricing options allow businesses to make the most out of their IT investments and includes on demand instances, dedicated hosts, per second billing, spot instances, reserved instances, saving plans and the free tier. Choosing the right pricing plan depends on the business needs and computational requirements.
- On-demand scalability within minutes
- 99% uptime commitment for all regions and availability zones
- A wide range of computing services including HPC, SAP, Windows and machine learning workloads
- Cost and performance optimization, 275 instances
- Processor, networking, OS, storage and purchase model options
- Minimized human error due to lockdown security models
- 89 compliance certifications and security standards including PCS-DSS, GDPR, HIPAA and FedRAMP
2. Microsoft Azure IaaS
Microsoft covers more regions than any other provider (54) and claims to be the largest public cloud services provider with massive 24 TB memory and fastest 160,000 IOPS single-disk storage. Azure IaaS offers over 90 public cloud compliance certifications, which is considered to be the most comprehensive any provider has to offer. Large enterprises can tap into virtually unlimited resources to manage their most resource-hungry apps such as AI and SAP HANA.
Microsoft’s IaaS infrastructure is designed for a wide range of workloads and allows businesses to easily scale their compute performance from 1 to 480 vCPU cores and 1GB to 24TB memory. Disk storage capacity ranges from 4GB to a massive 64TB, while networking speeds vary from 30Gbps to 100Gbps. Both Windows and Linux apps are supported along with a variety of programming languages and enterprise-grade apps, including Oracle, SAP, SharePoint and IBM.
The hybrid capabilities of Azure IaaS result in better app consistency, security and data management across cloud, on-premises and edge environments. Azure also claims to be the only public cloud solution that offers Backup-as-a-Service, which simplifies the process of protecting data from human errors and external threats like ransomware. Build-in control allows users to gain immediate insight into security and protect their hybrid cloud workloads.
Azure Cost Management comes free with all subscriptions and allows users to optimize their cloud spend. Different cost optimization options include transferring on-premises Windows/SQL Server licenses and paying in advance with Reserved VM instances for predictable workloads. This makes Azure IaaS a more cost effective option than AWS for Windows/SQL Server.
- 54 regions globally
- Up to 24 TB memory and 960 CPU threads
- 160,000 IOPS single disk storage and 3.7m IOPS local storage
- Up to 100Gbps hybrid networking
- 90+ compliance offerings
- 1 to 480 vCPU cores
- 4-64 GB disk storage
- 9%% single VM and 99.99% SLA across 2 or more zones
- Hybrid capabilities
- Azure Backup-as-a-Service
- Simplified data protection and security management
- Azure Security Center
- Governance tools, Azure Blueprints and Azure Policy make compliance and governance effective
- Free Azure Cost Management
- Azure Monitor provides a 360 view of network and app performance
- Transferable on-premises Windows/SQL Server licenses to the cloud
3. Google Compute Engine
Google’s configurable virtual machines are high-performance and scalable machines that run in Google’s data centers. Businesses can choose between general purposes, compute optimized and memory optimized VMs and seamlessly integrate them with other Google Cloud services including data analytics, artificial intelligence and machine learning.
The general purpose machines are available in three variants i.e.:
- N2 offer a good price-performance balance and can handle most general workloads such as web serving, business apps and databases
- N2D machines are suitable for workloads that require more memory bandwidth such as rendering, financial modeling and reservoir analysis
- E2 general purpose machines are suitable for businesses that want to run workloads that have no processor dependency and require up to 16 vCPUs. Workload examples include business critical apps, web servers, up to medium sized databases and other development environments
The compute optimized C2 machines are suitable for workloads that require high performance such as EDA, HPC and AAA gaming. Memory optimized machines are targeted at users who want to run workloads that require huge amounts of memory such as real-time analytics, in-memory caches and databases like SAP HANA.
Google Cloud Compute allows business to live-migrate host systems without having to reboot, which means they can keep their apps running during maintenance. Compared to regular VMs, preemptiable VMs allow running fault-tolerant workloads and batch jobs, which significantly reduces vCPU and memory costs without affecting performance.
The per-second pricing structure is based on how much businesses plan on using machines, persistent disks and selected resources. Different pricing options include VM instance, networking, sole-tenant, GPU and disk and image pricing. Google provides complete details of different pricing options so businesses can forecast expenses accurately and also offers a free trial to test drive its services.
- Pre-defined and custom VM configurations
- Time-based pricing
- Preemptiable virtual machines for short-term instances
- Live migration without requiring a reboot
- High-performance persistent disks
- Local, always-encrypted SSD storage
- GPU accelerators
- Global load balancing
- Supports Windows and Linux
- Committed-use discounts
- Dock container support using Google Kubernetes Engine
- Sustained-use automatic discounts
- VM instances reservations
- Optimization and right-sizing recommendations
4. IBM Cloud
IBM offers five cloud options i.e. public, private, public dedicated, multicloud and hybrid, allowing businesses to match their workloads with the most suitable cloud environment. Pricing options include free-tier, pay-as-you-go, reserved instances and subscription. IBM also offers cloud credits between USD 12k-120K for start-ups that meet the qualification criteria, while students and academic institutions (accredited) can get free access for non-commercial apps.
The most popular IBM cloud products include bare-metal servers, virtual servers, object storage, Kubernetes service, IBM Cloudant (NoSQL JSON), Blockchain platform, Watson Assistant and natural language understanding. IBM Cloud works great for enterprises that need a comprehensive set of capabilities for their multicloud, public cloud and hybrid environments.
The comprehensive range of IBM Cloud products include:
- Compute: bare metal servers, virtual servers, Red Hat OpenShift, Kubernetes, HPCaaS, Cloud registry, Cloud Foundry, Cloud Functions, IBM Power Systems Virtual Servers
- Network: Network appliances, domain name services, CDN, Direct Link, network security, load balancer, internet services and virtual private cloud
- Storage: Object storage, block storage, file storage, cloud backup
- Cloud Paks for multicloud management, integration, automation, applications and data
- Management: application manager, multicloud management, Netcool Operations Insight, cloud app management, availability monitoring, cloud automation manager, cloud event management, LogDNA (log analysis), Sysdig cloud monitoring, Istio and IT operations management
- Security: cloud activity tracker, cloud app ID, network security, SSL certificated, cloud certificate manager, hardware security module, hyper protect services, data shield, security advisor and key protect
- Databases: IBM Cloudant (NoSQL JSON), dbs for PostgreSQL/MongoDB, Elasticsearch, etcd and Redis and SQL Query, Db2 on Cloud, Db2 Hosted Flex and Informix on Cloud
- Anaytics: Analytics Engine, Apache Spark, SQL Query, decision optimization, Db2 warehouse, master data management, InfoSphere information server, streaming analytics
- IBM Watson: Watson Studio, knowledge catalog/studio, Watson Assistant, Discovery, IoT platform, text-to-speech, speech-to-text, language translator, natural language classifier, natural language understanding, visual recognition, tone analyzer, personality insights, machine learning, OpenScale, discovery for Salesforce and deep learning. Watson IoT platform connects devices, networks and gateways
- Mobile: Push notifications, Cloud ID and mobile foundation
- Developer tools: Cloud continuous delivery, Cloud CLI, starter kits, developer console for Apple, Cloud Shell, Tekton
- IBM Blockchain platform
- Migration: Mass data migration, SAP solutions, WebSphere on cloud
- VMware: cloud for VMware, Skytap, Veeam, Zerto and Horizon
- Open standards based
- Hybrid multicloud approach
- Untethered hybrid or private cloud
- Multivendor portability
- Multicloud management
- A comprehensive set of cloud products
- Industry-specific solutions
- 60 data-centers worldwide
5. Oracle Cloud Generation 2
Oracle offers a wide range of cloud infrastructure products designed for mission-critical apps. OC supports both modern development tools as well as legacy workloads, bridging the gap between past and the future. Oracle is also the only provider to offer a solution that can run Oracle Autonomous database so the choice is pretty obvious for businesses who want to run OA db.
From business analytics to app development, AI and blockchain, Oracle offers a comprehensive set of cloud products. The products are designed for businesses that want to run big workloads, high level of scalability and cost effectiveness. Most cloud services providers focus on uptime in their SLAs, but Oracle also covers manageability and performance in addition to uptime. The cloud products have built-in security with a layered approach to effectively respond to threats.
Businesses don’t have to rebuild everything from the scratch, thanks to Oracle’s commitment to open source and being a Platinum member of CNCF (Cloud Native Computing Foundation). Interoperable standards such as Kubernetes, Terraform and Helm as well as Oracle’s software development kits allow businesses to make the most out of multiple clouds and hybrid environments.
Oracle’s cloud offerings include:
- Compute: bare metal compute, virtual machines, container engine for Kubernetes, container registry and NVidia-based GPU offerings
- Storage: Object storage, file storage, local NVMe SSD, block volumes, archive storage, storage gateway and data transfer
- Networking: virtual cloud network, service gateway, load balancing, FastConnect, DNS, traffic management, health checks, email delivery
- Security and compliance: identity/asset management, key management, WAF, identity cloud service, cloud access security broker, database security
- Database management: Autonomous data warehouse, autonomous transaction processing, bare metal, virtual machine, exadata cloud service, and cloud at customer, NoSQL
- Analytics and big data: Oracle Data Science and Analytics Cloud, Essbase, big data service, SQL cloud, infrastructure data catalog and data flow, infrastructure streaming
- Management and governance: Monitoring, notifications, resource manager, tagging, cost management, audit, IT infrastructure and monitoring, application performance monitoring, IT analytics, log analysis
- Integration: API gateway, data integrator, GoldenGate, SOA cloud service and Oracle integration
- App development: Kubernetes container engine, container registry, hosted team development, Functions as a Service, fully-managed events service, Java, fully-managed API gateway, Mobile Hub, visual builder, digital assistant, data science, blockchain platform and Oracle content and experience
- Oracle Autonomous Database support
- SLA covers manageability and performance in addition to uptime
- Built-in security
- Open source standards support
- A wide range of services, including compute, storage, networking and database
- Compartmentalization and grouping, isolated sub-clouds
- Isolated network virtualization
- Support for both modern and legacy apps
- Bare-metal instances with 51.2 TB NVMe SSD
- No charges for outbound bandwidth up to 10TB/month
- Zero-trust architecture, encryption everywhere
- HTTPS-based REST APIs
- Native support for cloud-init and Terraform