There was once a time long long ago when telecommunications and computing were two different concepts. That was the age when phone company operators manually switched calls and computers like ENIAC were programmed by patches and cables. Since then the two fields have been on a convergence path. The many advances in both fields since the 1940’s make for exciting reading for computer and telecom fans, but rather than recount those achievements here I’d rather talk about a more modern achievement of note, the establishment of the Advanced Telecommunications Computing Architecture (ATCA or AdvancedTCA).
ATCA is an open standard that has been around since about 2003. It has been continually enhanced and today it is perhaps the most broadly accepted standard in the telecom industry, with over 100 companies
participating in development and implementation of the specification. Perhaps more important is the adoption of the standard in the telecommunications industry. A review of wikipedia entries and other open info (like the Intel Embedded and Communciations Alliance) indicates typical “hockey-stick” implementation seen in other highly reliable, highly virtuous standards. IDC projects the ATCA market will be about $2.7 billion in size by 2013. I think the global financial crisis and the ongoing wave of mergers and purchases of smaller comms and equipment providers by larger ones will accelerate this trend even faster, as the need for modular low cost, highly reliable standards is needed even more.
Network equipment providers face two challenges that they are addressing with ATCA: 1) the need to continue to deliver new platforms and applications and, 2) the need to reduce costs and improve productivity. ATCA provides a great opportunity to address these needs. ATCA standards provide a common platform which provides lower cost, reduced maintenance, the ability to use third party boards, and the ability to reduce vendor lock-in (more on ATCA capabilities is below).
In my opinion, enterprise CTOs should work to accelerate moving the ATCA standard and compliant products into data centers. It results in more computer power per square inch, higher reliability, power savings, cost savings, long term maintainability, and a path for upgrade that does not require forklifts. ATCA is not something that currently scales down to small network devices, but it is something that I believe will prove to be perfect for data center server support.
Here is more on ATCA:
- Boards (blades) in an ATCA shelf are hot swapable.
- There is not a “bus” for communications in an ATCA shelf. Instead, boards communicate point to point, which is faster and ensures there is not a single point of failure like in the bus model.
- Any switching fabric can be used.
- Boards can be processors, switches or specially designed advanced cards, if desired.
- The most advanced shelf management capability ever designed is in the ATCA container. If any sensor reports a problem the shelf manager can take action or report the problem to a system manager. This action could be things like turning up a fan or powering off a component or telling a human that something needs to be replaced before failing.
- It is designed for very high reliability and very high availability.
- It runs cooler, even with its higher powered processors.
- It supports a healthy multi-vendor, interoperable ecosystem.
- It is based on open standards vice proprietary (locked-in) solutions.
Now back to the opening idea of this post. Telecom and data and compute power are not separate things anymore. Each are closely interwoven and successes in one thrust can make a huge positive difference in capabilities in other areas. As organizations and users grow more accustom to the power of cloud computing they will demand higher and higher levels of reliability and resiliency from their server providers. And as service providers provider higher levels of reliability and throughput cloud compute providers will see more and more success which will place increased requirements on their capability. In both cases, ATCA will provide the agility, resiliency and reliability required, which will drive its adoption further and further into the telecon and data worlds.