With all of the buzz today around “the cloud,” it is important to remember that we are not discussing a monolith. Rather, there are different types of cloud deployment and applications within the cloud ecosystem. While it is true that 83% of enterprises will use the cloud by 2020, the type of cloud deployed and the ways in which organizations engage with the cloud may vary greatly.
Now that we know where the cloud came from, it's important to understand the types of cloud and the suite of solutions within the cloud. Let's unpack the four main deployment models and the three main service offerings across the cloud deployment models.
The most common flavor: public cloud
The public cloud is what people most often think about when the topic of cloud computing arises. These solutions are readily available from Google, Amazon, Microsoft, and others. Amazon Web Services alone has over a million active users.
In the public cloud, all computing infrastructure is located in the data center(s) of a cloud service provider and each user has a virtualized private environment within the larger public ecosystem. The provider has physical control of the hardware and is responsible for all aspects of data security, IT management, and support.
The model generally applied by public providers of cloud services is one of shared resources, offering dramatic scalability and elasticity at a fraction of the cost to end users compared to maintaining the hardware and personnel on premise. This is the main model offered by AWS and Microsoft Azure.
Taking a baby step into cloud: private cloud
For users who are uncomfortable with handing over all control and support of their IT infrastructure or data to a third party, there is a private cloud option. For a higher cost, an enterprise can set up a dedicated and private network located either on-premise or off. Here, there is an ability to privately share virtualized resources with heightened security protocols or obligations.
The tradeoff is a relative increase of cost to stand up and maintain the environment and less elastic scalability because those using a private cloud are still dependent on their own hardware scale and personnel.
Somewhere in the middle: hybrid or community clouds
For users looking to gain the benefit of cost savings and scalability without fully relinquishing oversight of their data, there is the hybrid cloud option. In this model, an enterprise would have both a private and public cloud infrastructure that would interact.
An alternative approach to cloud deployment is a shared private cloud or community cloud. Here, several organizations pool resources and share data management responsibility. This mitigates some of the access and control concerns organizations may have without losing the cost and scale benefits of cloud computing.
How enterprises engage with the cloud
The other way that people differentiate the cloud is in the model of services employed. Not every enterprise uses the cloud in the same way, so understanding the various models is helpful in architecting your cloud ecosystem.
SaaS - Software as a Service
SaaS solutions are familiar to most users in the ediscovery space: a completed product or tool is provided via web access to an end user. The tool or software is completely managed and supported by the provider. The end user no longer has to worry about data security, infrastructure, or overall maintenance of the tool and can just focus on the ways they will use the software.
With SaaS, users simply need to log in and they are ready to go, without the need for middleware or infrastructure concerns. Common examples of SaaS solutions that most of us engage with daily include web-based email, CRM tools like Salesforce, Google G Suite, and collaboration tools like Slack.
IaaS - Infrastructure as a Service
With IaaS, an organization procures the building blocks for its IT ecosystem from a third party. It is then provided in a virtualized environment accessible via the internet. In this scenario, a cloud service provider like Amazon Web Services (AWS) hosts and manages the components that comprise a traditional IT infrastructure including servers, network and storage hardware, and virtualization (hypervisor) layer. In addition to the raw components of the infrastructure, a cloud service provider also is responsible for maintaining data security oversight, load balancing and optimization, disaster recovery capabilities, and redundancy. Generally, the end user has to install their own virtual machines, operating systems, and support applications and manage their own configurations.
Increasingly, even large enterprises or law firms are turning to IaaS solutions to increase their scalability, adaptability and to address the evolving cyber threat landscape. The cost of continuing to scale internal IT ecosystems to handle the volume, variety, and velocity of data combined with the increasingly sophisticated cyber threat landscape is driving up overall IT costs to the point where some organizations are unable to keep up with a standard on-prem model.
A major appeal of the IaaS model is that end users pay on an as-used basis, with costs scaling up only when the workload increases and not in anticipation of an increased workload.
PaaS - Platform as a Service
PaaS is a combination of the previous two models, with a third-party provider delivering a combination of hardware and software tools to end users via the internet. Because the third-party hosts both the software and hardware internally, the end user no longer has to install hardware or software behind their firewall in order to host applications or develop. Like the prior two models, the benefit here is that it frees up the end user to create and develop their own tools in lieu of focusing on infrastructure and underlying services.
As with IaaS, a benefit of PaaS is that users only pay for as much as they use, generally calculated on a per-user basis. Unlike with IaaS, the service provider for a PaaS offering will install the relevant operating system and middleware to support the end user.
What does this mean for ediscovery?
As an end user in ediscovery, SaaS and IaaS are the most common iterations of cloud-based services we engage with. Anytime we access review platforms via the internet as an end user, we are effectively using a SaaS solution. Even the most staunchly anti-cloud buyer, ironically, is already engaging with this cloud model even if they do not know it.
Alternatively, IaaS is something that is newer to the ediscovery universe. This cloud computing model is seen either when end users directly engage with a cloud provider like AWS to support an internal enterprise or law firm ediscovery program or when the technology partner an end user has selected has an IaaS deployment.
Some technology partners take advantage of both cloud deployment models, internally leveraging IaaS to increase scalability and adaptability while offering their end users a simple streamlined SaaS option. This best of both worlds approach is often what is meant when a technology provider is described as a cloud provider.
The future of cloud in ediscovery
With ediscovery practitioners facing rapidly expanding data varieties and volumes, coupled with the need to reduce time to insight without dramatically increasing cost, cloud adoption is not a question of if but when. The cost of keeping pace for the average enterprise or law firm is too great, and the risk of falling behind too great to continue with the status quo.
Additionally, the rapidly scalable computational power offered by cloud-based systems purpose-built to capitalize on the decentralized elastic nature of the cloud enables the development of more robust and powerful analytic engines. The need to get through massive data volumes and get insights faster is another key driver to the cloud that I foresee increasing adoption in the coming months.
Who's Afraid of the Big Bad Cloud blog series:
- What is cloud and where did it come from?
- Why cloud adoption has reached a tipping point
- Understanding why the cloud isn't actually scary