Oracle Training in Velachery

Oracle 2020

A Glimpse Into the Future of Database Management in Oracle

• The year is 2020, and the roles and responsibilities of the Oracle professional have changed dramatically over the past 15 years.
• The Bureau of Labor Statistics shows the Oracle DBA job as top 10 job for growth
• To fully understand the benefits of computer hardware in the year 2020, we must begin by seeing how the constant changes in CPU, RAM, and disk technology have effected database management over the past six decades.
• Once we see the history in its correct perspective, we can understand the evolution of Oracle database systems.

Oracle Hardware Advances Between 2005 and 2020

• Gordon Moore, Director of the Research and Development Laboratories at Fairchild Semiconductor, published a research paper titled ?Cramming More Components into Integrated Circuits in 1965.
• Moore performed a linear regression on the rate of change in server processing speed and costs, and noted an exponential growth in processing power and an exponential reduction of processing costs. This discovery led to Moore's Law, which postulated that CPU power gets four-times faster every three years.
• However, the real? Moore's Law cannot be boiled down into a one-size-fits-all statement to the effect that everything always gets faster and cheaper.
• Prices are always falling, but there are important exceptions to Moore?s Law, especially with regard to disk and RAM technology.
• As we can see, these speed curves are not linear, and this trend has a profound impact on the performance of Oracle databases. Let?s take a closer look.

Disk Storage Changes

• I'm old enough to remember when punched cards were the prominent data storage device.
• Every year, I would get my income tax refund check on a punched card, and we would make Christmas trees from punched cards in the Data Processing department.
• My college kids have no idea what the term ?Do not fold, spindle or mutilate means, and they missed out on the fun of dropping their card deck on the floor and having to use the giant collating machines to re-sequence their deck.
• In 1985, I remember buying a 1.2 gigabyte disk (the IBM-3380 disk) for more than $250,000. Today, you can buy 100 GB disks for $10, and 100 GB of RAM for $100. With these types of advances, Moore's Law for storage.
• Disk storage costs fall 10x every year.
  Storage media is obsolesced every 25 years.
• Note that the change to Moore?s Law for disks shows the limitations of the spinning platter technology.
• Platters can only spin so fast without becoming aerodynamic, and the disk vendors were hard-pressed to keep their technology improving in speed.
• Their solution was to add a RAM front-end to their disk arrays and sophisticated, asynchronous read-write software to provide the illusion of faster hardware performance.

RAM Storage Changes

• Today in 2020, you can buy 100 GB of RAM for only $100, with access times 600,000 greater than the ancient spinning disk platter of the 20th century.
• In 2020, a terabyte of RAM costs less than $200.
• The introduction of Quantum-state Gallium Arsenide RAM in 2009 was the largest breakthrough in RAM in more then 40 years. Before 2009, RAM always became cheaper every year, but it did not get faster.
• This meant that CPU speed continued to outpace memory speed, and RAM subsystems had to be localized to keep the CPUs running at full capacity.
• Until 2009, RAM speed remained constant at about 20 microseconds (millionths of a second), and even the solid-state database had to deal with the continued increasing speed of CPU resources. Let?s examine the CPU changes over the past 15 years.

Processor Changes

• The same trend also exists for processor costs and speed. In the 1970s, a 4-way SMP processor costs over $3,000,000.
• Today in 2020, the same CPU can be purchased for under $300.
• CPUs continue to increase speed by four times as much every three years and cut cost in half.

I/O bandwidth capacity doubles every ten years with Oracle

• These super-cheap, super-fast processors sounded the death-knell for the age of small computers, and server blades (and Oracle10g Grid computing) were replaced by large, monolithic servers.
• Between 2005 and 2009, RAM had to be physically localized near the CPU to keep the processors running at full capacity.
• After 2009, the speed of RAM increased to picoseconds (billionths of a second); this development changed server architectures.
• The largest source of latency was not the wires between the CPU and RAM, and fiber optic cables were required to keep up with the processing speeds.
• During this period, computer servers first began to take on the familiar tower configuration that we know today.
• As we all know, the tower configuration is required to minimize the fiber optical length between the CPU and RAM, and this is required to keep the CPUs operating at full capacity.
• As RAM speed broke the picosecond threshold and approached the speed of light, even the fastest 20th century networks could not keep up with the processing demands.
• Quantum mechanics and atom-state technology were combined with fiber optics to improve line speeds to keep pace with the hardware.
• These advances in hardware made mini-computers instantly obsolete, and management recognized that multiple servers were far too labor intensive.
• Starting in 2005, we began to see the first wave of the massive server consolidation movement.
• The large, 64-bit servers with 16, 32, and 64 CPUs became so affordable that companies abandoned their server farms in favor of a single-server source.