About Ciscos blades
Some people think, that Ciscos move to develop blade server is Ciscos next step to take a large dent out of the server market in addition to it´s networking share. Well … i would say: It´s one of Ciscos strategies to survive at all. Switching starts to get a commodidity market. You can get a 5 port 100 MBit/s el-cheapo switch for 9 Euros. A ProCurve Switch 1700-8 costs less than 50 Euros and that´s brand name managed Fast Ethernet-Switch. And now there are even competitors trying to make 10 GBe an commodity like Andy´s Arista Networking. I assume in 2 years from now you can buy standard silicon for 10 GBe/s at prices like 1 GBe switching today. So i don´t really think that switching will be really one of the core markets of Cisco in the future. What is a router? Essentially an custom build embedded server with some networking hardware and a large bunch of software. The key differentiator of Cisco is the large bunch of software … the IOS. But now let´s assume there is a general purpose operating system with networking capabilities nearly as good as a custom build operating system. It´s only a matter of time that a vendor takes standard components to build a router out of it. In earlier times, the limited I/O resources of standard hardware were a problem but with PCIe this isn´t really a problem. Cisco will have the same problem like high-end server vendors since the dot.com bust and the high-end storage vendors at the moment in the next few years. If they are concentrating on stuff that will be commoditized soon, they would be essentially toast in a few years. That´s something server vendors including Sun had learned on the hard way. The move of Cisco to develop blades was a necessary step to protect at least a bit of the market share in networking. They had to come out of their networking corner. They have to give a compelling answer to the question: “Why should i buy equipment from Cisco when there is cheaper oder better equipment with the same quality ?”. Integration of different technologies may be the answer. But it will need more than just a blade rack with a nicely integrated FCoE enabled switch and some virtualisation to get a major player in this area. In many medium to large companies networking and server administration are very different departments with quite different vendor relations. That will be a tough sell for Cisco sales people. I´m really curious about the stuff Cisco will announce in the upcoming time. I hope there is a more compelling story than just “We have blades now, too … with a Cisco switch”. On the other side: There is money to be made in commoditizing away the markets of midrange and high-end networking gear, thus this opens opportunities for other vendors who haven´t to protect an exisiting business model. At the end the move of Cisco is perhaps a perfect validation of the points Jonathan made in his recent blog article.