- The cloud instance IP can change In a shared network after you stop or delete the instance. Therefore, you need to keep your router or any applications connecting to the instance updating the IP after these operations.
- IP network allows you to isolate the private network from the public Internet.
In the shared network, each instance is assigned a private IP address from the common pool of Oracle provided IP addresses. In an IP network, you can define an IP subnet in your account. The address range of the IP network is defined by the IP address prefix defined when creating the IP network. The following diagram from Oracle Documentation  explains the two network offerings:
The main differences between the two network options are:
For years, we have been used to the block storage, including SAN, iSCSI, and local disks, to keep our data. Block storage provides fix-sized raw storage capacity where each storage volume is an independent desk managed by the operating system. The data stored in the block storage don't have any concept of data type and format. Instead, block storage relies on applications or file systems to keep track of data location, format and meaning. The block storage options in the cloud include:
Object storage stores data as an object with ID, metadata, and data. You can directly access each object without going through disk scans or directories. The block storage options in the cloud include:
Oracle Bare Metal Cloud Service (OBMCS) is Oracle's next generation Infrastructure as a Service (IaaS). It was announced at OpenWorld 2016 and now offers compute, storage, network, governance, database, and load balancing services. Oracle Bare Metal Cloud Service is claimed to provides high performance, better network management, flexibility and lower cost. The new IaaS platform starts a close competition with Amazon AWS. In this blog, let's look at its network architecture.
The network architecture of Oracle Bare Metal Cloud Service is similar to Amazon AWS. It is built on the virtual network concept. In Oracle Bare Metal Cloud Service, the virtual network is called the Virtual Cloud Network (VCN). The network is comprised of Region, Availability Domain, Subnet, Routing Table, Internet Gateway, and Dynamic Routing Gateway. You can define a VCN and assign both private and public IP to your cloud instances. The following slide provides an excellent summary of the fundamental concept and the architecture.
I also include a slide for Amazon Virtual Private Cloud  to hep understand the related the concepts.
With Oracle Bare Metal Cloud Service, Oracle Public Cloud now offers much better security and network management.
Created 3/23/2017, Last Updated 6/28/2017
Please refer to the following white paper for the details:
This is a question that has been asked by most people who are new to the cloud. It's also a major reason making people feel hesitant to moving to the cloud. Let's talk about it today.
The first answer to this question is a question. Do you want to provide your service over the Internet? In the earlier discussion of Why Choose Cloud Computing? I discussed some major reasons that makes the cloud solutions attractive. If there are compelling reasons to leverage those benefits, then the following conversation makes sense. If not, you can stop here and find something else to read.
If the Internet access becomes necessary for your service, then you have to host it in the cloud. You then need to consider what level of security you'd like to provide for your cloud service and how much investment you want to put into this. To help you understand this, I divide the cloud security into several levels.
Level 1: Protecting the Data.
Because the service is published online and the data is sent over on the Internet, the first thing you need is to safeguard the data. The followings are several important considerations:
Level 2: Restricting the Access.
On this level, you would identify and decide who from the Internet can access your service and what they can access.
Level 3: Cleaning up your endpoints.
This is the hardest and most vulnerable part because of the Internet access. Your users are not security experts. They may not be aware of Malware (or malicious applications) running on their machine or mobile phones or the network connections that is open to others. Then, your service needs to help them detect the malwares , identify suspicious activities and inform anything that is threatening.
Level 4: Auditing the service
Another level of protections is a detailed auditing on who has accessed on what data and when. A comprehensive report and analysis of the service usage are another level of security protection, which allows you to discover issues and respond to the threats quickly.
Because of the Internet exposure, it's natural to have concerns about the cloud security in the first place. However, security is not a new topic to IT services. You just evaluate the gains, impacts, and challenges, and then choose a cloud platform to meet your needs.
Created 9/23/2016, Last Updated 5/30/201
After my Oracle OpenWorld session on Oracle GoldenGate Cloud Service, Steve from the audience came to me and asked a simple yet excellent question: What is Cloud? I then realized that this basic concept needs an explanation. Let's talk about it here. A good understanding of this concept can help you understand the design and functionality of new cloud services.
In the context of Information Technology (IT), Cloud means the Cloud Computing, which refers to the IT resources and applications which you can access over the Internet. This is it! There no other requirements to call it cloud computing. However, there are some misconceptions, so let's clarify them:
Scaling is a major benefit of cloud computing. With scaling support, users don't have to guess their computing resources at the beginning. Instead, cloud services can increase or decrease the compute resources on demand . This blog: (1) explains what scaling is, (2) discusses the feature design, and (3) explores scaling support offered by cloud vendors.
1. What is Scaling?
Scaling means adding/reducing resources. There are two types of scaling :
For a cloud service providing the scaling support, the following issues need to be considered:
3. Example Scaling Support
Let's first look at Oracle Java Cloud Service. It is a service providing both scale up/down and scale in/out support . The scale up/down is provided as part of the node management. You can select the node and click on the Scale Up/Down menu. A scale up/down dialog allows you to specify the compute shape, additional storage, and where to add the additional storage. This scaling support requires service to be put into maintenance mode and restart. The scale in/out is also provided by allowing users to add nodes to the compute cluster. Some screenshots are listed here:
Second, let's look at AWS Auto Scaling . AWS Auto Scaling manages a pool of servers called the Auto Scaling Group. It can replace/grow/shrink the size of the pool based on the workload. Please refer to this video for a good demo of this feature.
In summary, the following steps are needed to set up an Auto Scaling in AWS:
Additionally, it's good practice to define an auto scaling group across different availability zones (AZ) to ensure high availability and to use the Auto scaling Cooldown feature to avoid turning down too many instances (by default 300 seconds). 
Clearly, a good cloud service should: (1) provide a scaling support so that the cloud resources can be adjusted when the resources are under stress or remain idle, (2) have the ability to be easily managed, and (3) allow nonstop services. As a cloud user, it's important for you to know that the scaling process can be complex and sometimes require you to use advanced offerings such as the Connection Draining support in AWS. It's important for you to understand your requirements and choose a plan that works for you.
Thanks for reading.
Created 3/30/2016, Last Updated 5/31/2017