This is the last newsletter in a series that has explored our premise that we are returning to a new era of mainframe computing. This newsletter will describe what is meant by a mainframe, provide an example of the emerging generation of mainframe computer and will talk about how some old friends don't like each other much these days.
One clear characteristic of a mainframe computer is scale – a mainframe is notably larger than other types of computers; i.e., servers and PCs. Because of their scale, for the last 40 years mainframes have supported virtualization. Mainframes can also add or hot swap system capacity non-disruptively and granularly, to a level of sophistication usually not found on most servers. Mainframes are designed to handle very high volume of input and output (I/O) and deal with massive databases and files.
So while we realize that today when someone uses the term mainframe, most IT people think of the IBM zSeries; e.g., System z9 and System z10. We certainly recognize those as being contemporary mainframe computers. However, when we look forward we think of what is being produced by the Virtual Computing Environment (VCE). The VCE is a coalition that is comprised of Cisco, EMC and VMware. One of the key concepts that underlies the VCE is the concept of Vblock solutions. Vblock solutions are pre-integrated and tested infrastructure packages that bring together technologies from Cisco, EMC and VMware. For example, mid to large sized Vblocks are based in part on Cisco's Unified Computing System (UCS).
A Vblock combines virtualization, computing, networking, storage, security, and management and the VCE promises to deliver them as a single entity with a single line of support. If you look inside of a Vblock you will see a lot of the characteristics of a mainframe computer. For example, consider scalability. A high end Vblock can support up to 6,000 virtual machines.
One way to look at this series of newsletters is that it demonstrated that yes indeed we are entering a new era of mainframe computing. This series showed how today's thin and zero clients are similar to the ADCII and 3270 terminals of the previous era of mainframe computing. The series also showed that the WAN that supports the emerging era of mainframe computing has to be notably different than the WAN that supported the previous era. It has to be orders of magnitude faster and has to support WAN optimization ubiquitously.
The series also pointed out that the next generation of mainframe may not come from a single vendor, but from an alliance. On the one hand, an alliance sounds appealing. The vendors can promise that the result of the alliance will be a solution that involves only the best-of-breed components. However, while it is relatively easy for companies to announce an alliance, it is much tougher to make them work long term. Another dynamic that impacts all IT organizations is that it is not exactly a secret that when Cisco first announced UCS over a year ago, it really angered its traditional partners IBM and HP. As a result, Cisco and the VCE is now going head to head with at least two other behemoths – IBM and HP. While not quite as large, it is difficult to overlook the behemoth that is Oracle and its feisty leader Larry Ellison.
However, another way to look at this series of newsletters is that it highlighted the fact that now, more than any time in the last two or three decades, so much of IT is in play. For example, in a recent newsletter we asked the question of how vulnerable was Cisco. We believe that Cisco is vulnerable, particularly in selling into small and midsized branch offices. We also believe that a huge battle is underway for the data center, a battle in which each behemoth has a lot to win and a lot to lose. The net result is that this is clearly a time for IT organizations to re-think their technology and vendor strategies.