Exadata and Exalogic received their names because of the exabytes of information that the data centers of the future will be required to handle. Our society is becoming increasingly inundated with digital information. Today, information is broadcast from satellites and transmitted over the airwaves, cables, fiber networks, and through other means. In 2004, monthly Internet traffic exceeded 1 exabyte, the equivalent of 1000 petabytes. In 2010, monthly Internet traffic exceeded 20 exabytes.
An exabyte is a unit of information or computer storage equal to 1 quintillion bytes. One kilobyte equals 1000 bytes. One megabyte equals 1000 kilobytes. One gigabyte equals 1000 megabytes. One terabyte equals 1000 gigabytes. One petabyte equals 1000 terabytes. One Exabyte equals 1000 petabytes. See Table 1-1 for a visual representation of the size of information from byte to exabyte.
Integrated Machines vs. Appliances
Some of the characteristics of engineered solutions, such as the preconfigured assembly of components, are similar to appliances. Appliances do not have configuration options and are designed to perform a simple task, such as a toaster that heats bread to a certain temperature. Exadata and Exalogic are not true "appliances," however. Instead, they are engineered platforms that have been optimized to run application servers and databases. Hence, many of the traditional rules for managing application servers and databases continue to apply.
In the "One Day Installation Challenge" video on YouTube, Oracle demonstrates how to install an Exalogic machine in less than ten hours. In contrast, a company following the do-it-yourself platform development approach described earlier might spend more than six months to integrate and deploy a comparable platform.
Read more of this book
Excerpted from Oracle Exalogic Elastic Cloud Handbook (McGraw-Hill; 2012). Written by Tom Plunkett, TJ Palazzolo and Tejas Joshi, with permission from McGraw-Hill. Read the entire chapter here. To purchase a copy of this book, please visit Amazon.com.
With traditional IT systems, when a customer calls the support desk, before a problem can be diagnosed, a support team asks a customer numerous questions. These questions ask about the specific component vendors and versions used in the host environment, including operating system, hardware, network, storage, firmware, and patch levels. All must be determined before the actual process of diagnosing the problem can begin. Furthermore, the customer support organization for the specific component will not typically have access to an identically configured environment, especially if some of the elements are supplied by a different vendor. There are limitless combinations of how such components can be configured together. Indeed, during the issue resolution process, it may transpire that the root cause is actually related to a different element of the system belonging to a different vendor. When this occurs, the customer will need to begin this whole support issue process again, with the other product vendor, thus prolonging the issue resolution time.
In contrast, with integrated machines, you simply tell the customer support person the machine model you are using. Furthermore, the customer support desk has access to an identically configured machine, making it easy to replicate the problem that the customer has encountered and dramatically simplifying the troubleshooting process.
Oracle Exadata Database Machine
Although the focus of this book is the Exalogic machine, you'll also find it useful to have a basic understanding of its close cousin, Exadata. There is value in comparing and contrasting their architectures, and you need to keep in mind that both will often be used together.
More information on Oracle's Exalogic Elastic Cloud
Read about the big 2010 reveal of the elastic "cloud in a box"
Peruse this guide to Oracle cloud computing
The Oracle Exadata Database Machine is an integrated software and hardware platform that has been designed for data warehouses, Online Transaction Processing (OLTP), and database consolidation. The Exadata machine's InfiniBand fabric enables extremely fast input/output (I/O) communication between the storage server and the database server. (InfiniBand is also used within Exalogic for high performance I/O and is described in detail in Chapter 2.) Oracle Exadata includes intelligent storage server software that considerably increases the performance of data warehouses by reducing the amount of information that needs to be communicated between the data warehouse server and the storage server.
Exadata uses dedicated servers, or nodes, for storage and processing. The ratio of storage nodes to compute nodes in the Exadata machine is based on an understanding of the business problems that Oracle's customers are attempting to solve. (Compute nodes for Exalogic will be described in detail in Chapter 2.) The Exadata machine that is intended for data warehouses has a ratio that differs from the Exadata machine that is intended for OLTP applications. Both types of Exadata machines rely on the InfiniBand fabric to enhance performance between the database server and the storage server.
The database servers can utilize Oracle Linux or Oracle Solaris as the operating system. (Both operating systems are available for Exalogic, and are described in Chapter 2.) The storage servers use only Linux, and the ZFS Storage Appliance runs Solaris. Oracle 11g Release 2 is the current version of the Oracle database that runs on Exadata.
Appliances and similar solutions have been involved with data warehousing for a long time. Exadata, however, has optimizations that existing data warehouse appliances do not have. Furthermore, Exadata also provides the first optimized solutions for OLTP.
Exadata performance enhancements have focused on eliminating unnecessary communication between the database server and the storage server and speeding up communication between the database server and the storage server.
Oracle Exalogic Elastic Cloud
Oracle Exalogic Elastic Cloud is an engineered hardware and software machine designed to provide a platform for a customer's entire application portfolio. Exalogic includes hardware connected by an internal InfiniBand network fabric. An Exalogic machine can be connected with additional Exadata and Exalogic machines. Oracle's market-leading Java Enterprise Edition (Java EE) application server, WebLogic Server, has been reengineered for deployment to Exalogic and uses specific performance enhancements when running on Exalogic. The InfiniBand network fabric offers extremely high bandwidth and low latency, which provides major performance gains with respect to communication between the application server and the database server, and with respect to communication between different application server instances running with in the Exalogic system. Physically, Oracle Exalogic Elastic Cloud can be viewed as a rack of physical server machines plus centralized storage, which all have been designed together to cater to typical high-performance Java application use cases. Figure 1-2 shows the front view of an Exalogic machine.
In summary, Exadata is the database machine for the data tier and Exalogic is the middleware machine for the application tier.
Why Would an Organization Adopt Exalogic?
Why would an organization adopt the Oracle Exalogic Elastic Cloud machine? Just because Exalogic is an extremely high-performing Java solution, it does not mean that it is the perfect solution for every organization. Some organizations have architectural considerations that counsel adoption of alternative solutions. Further chapters will dive into these architectural considerations in greater detail, but the next few paragraphs describe some of the areas for which customers may find great value in adopting Oracle Exalogic Elastic Cloud. The next step is to understand Exalogic's hardware architecture in greater detail.
ABOUT THE AUTHORS
Tom Plunkett is a senior consultant with Oracle. He has taught graduate-level computer science courses for Virginia Tech's computer science department and has authored several books on topics related to Big Data, Cloud Computing, Java, and Service-Oriented Architecture. Tom has spoken internationally at more than 30 conferences and delivered hundreds of presentations worldwide on the subjects of Big Data, Cloud Computing, Java, and Service-Oriented Architecture.
TJ Palazzolo is a senior principal curriculum developer with Oracle University. He is the lead curriculum developer for Exalogic Elastic Cloud, WebLogic Server, and Oracle's virtualization products. TJ develops classroom and online training materials on these and other Oracle Fusion Middleware products. He is also responsible for maintaining the Exalogic rack installations used by Oracle University to deliver education.
Tejas Joshi is a technical director with Oracle Consulting in the UK. He has worked on many high-profile multi-million-dollar solutions during the past fourteen years in both the public and private sectors, drawing recognition through awards including Consultant of the Year 2012 for Western Continental Europe. Tejas is a frequent speaker at events such as Cloud Academy, Oracle Services events, Oracle networking events, and Oracle product launch events.
This was first published in September 2012