As the open source faithful have gathered in Portland for the annual O’Reilly Open Source Convention (OSCON) Canonical, the company behind an ever popular Linux distribution Ubuntu, announced that it has paired up the latest version of the free DB2 Express-C v9.7 with the latest version of Ubuntu 10.04 LTS to deliver a free enterprise-grade cloud database appliance.
Database appliances, especially for data warehousing, are all the rage. These database appliances deliver a package of hardware and database software that are put together and optimized by the DBMS vendor. Most of the industry analysts have placed their bets on continual growth of “database appliance”. Netezza, Greenplum (now EMC), Kickfire are just a few examples of pure database appliance companies. In addition, major DBMS vendors like Oracle, Teradata and IBM also offer database appliances.
The main benefit of using a database appliance is balanced approach to performance and, most important, reduced bring-up and ongoing operating costs. The main drawback is a large initial capital outlay as you are going to get a system that is sized for anticipated needs. Appliance vendors such as IBM try to mitigate this issue by delivering modular design that allows for incremental growth. While this does help, most customers are very much attracted to the “pay-as-you-go” model with zero … yes zero … capital expense. “No CAPEX (capital expense)” is the mantra for the Cloud Computing Infrastructure as a Service (IaaS) and it is getting attention of CFOs and CIOs in every industry and every company size.
This is where Canonical announcement becomes interesting. What Canonical did, with help from us in the IBM DB2 team, is put out a database appliance that anybody can benefit from, not just large corporations with big capital budgets. The entire software stack in the appliance is free (both Ubuntu and DB2 Express-C are free products) and the hardware can be rented from Amazon EC2 by the hour. Renting hardware from Amazon EC2 means that you do not have any upfront expense. You get a bill at the end of the month and you pay only for the resources you actually used and only for the time that your servers were powered on.
At this point in time, you should be thinking: “this sounds to good to be true”, and you are probably thinking “what about support, security, privacy etc.?” That is good thinking! Let me start by saying that you are not going to get an equivalent of the multi-million dollar Oracle Exadata or the IBM Smart Analytics System for a few cents an hour with this solution. However, you may be pleasantly surprised just how much value you actually do get for a free or a few cents an hour.
Let’s start with support. The operating system in this appliance is Ubuntu Server 10.04 LTS. “LTS” stands for Long Term Support. Which means that support and updates for it will be available for the next 5 years. While the free DB2 Express-C does not come with IBM support (it is free after all), a low cost yearly subscription is available that provides full support from IBM in addition to some very handy extra features and functions. Another important point to mention is that DB2 Express-C is not some imitation of DB2 or a crippled version of it. It is a real DB2 built from exactly the same code base as all other editions of DB2. It is enterprise grade DBMS through and through. For those that get the yearly subscription, support is delivered by exactly the same support teams around the world that service and support DB2 Enterprise Edition for some of he world’s largest companies.
Security and privacy is one of the first concerns that come up whenever Cloud Computing is discussed. People are concerned about placing their data in to a data center that they do not control and having access to this data over a public network infrastructure. These are valid concerns, however, they can be easily addressed by employing isolated systems and secured private communication. Amazon EC2, for example, offers a really interesting service they call Virtual Private Cloud or VPC. In a nutshell, VPC allows you to rent compute and storage capacity from Amazon yet have these systems completely isolated from the internet and accessible only from your data center through a secure Virtual Private Network (VPN) channel. The same VPN that companies trust to provide secure access for their mobile employees. In essence, VPC allows you to treat resources in the Amazon data center as an extension of your own data center secured and protected by the same security procedures and the same IT security team that keeps evil doers away from the systems and data in your own IT infrastructure.
For some people, using systems in someone else data center is not acceptable for compliance reasons. Swiss financial industry regulations, for example, forbid placement of certain kinds of data in a data center located beyond its borders. There are also governmental and industry regulations that require that auditors be provided with physical access to the IT equipment. In these cases, customers may still want to get the benefit of having a dynamically provisioned shared IT infrastructure but have it housed in their own data center. People call this “Private Cloud”; I call it “Good Idea”. There are a number of products on the market that help people build their own private clouds. Ubuntu Server 10.04 is one such product. They call it Ubuntu Enterprise Cloud or UEC for short. UEC is based on Eucalyptus. Eucalyptus is an open source clone of the Amazon EC2 and as such, provides a very easy way to build an Amazon EC2 API-compatible private cloud infrastructure.
The Cloud Database Appliance from Canonical looks like a pretty good choice for getting the benefit of a database appliance without prohibitively high capital expense. Build on an enterprise-grade software stack, it can be deployed on a public cloud infrastructure and can support both customer facing and private enterprise applications in a very secure scalable and highly available environment. When deployed on a public cloud infrastructure, the only cost is for the rental of the hardware resources and these costs are incurred only when the system is powered up and you pay only for the resources that you actually used. Last but not least, with many customers building their own private clouds, Canonical Cloud Database Appliance based on DB2 and Ubuntu will present a very interesting option for creating free or very low cost development, test and even production database environments.
What do you think?