NO MORE UPDATES TO THIS PAGE PLEASE. SUBMIT ALL FUTURE COMMENTS TO

Remote access to PCs (pg 38)

Some vendors (eg HP, IBM) have hardware cards to allow remote control of the machine. HPs was called RIB (Remote Insight Board) and would allow SNMP and remote console access their ProLiant^? servers. This was one way their server-class PC machines differed from other PCs.

[ Tom's reply: Noted. Status: DONE ]

2.3 Many Inexpensive Workstations (pg 44)

Add a new section on blades. Discuss difference between diskless and diskful blades, and how this changes how they are managed (blade explodes on a diskless blade, just bring up another and it gets all the data from the SAN; blade explodes with a disk present requires an automated server rebuild). Important to note the impact high-density blade farms have on heat and power requirements. Much more than you ever expect!

Warning: treat blades as black boxes, not as seperate Unix machines otherwise this could happen to you:

Anecdote: A division of a large multinational was planning on replacing their aging multi-CPU server with a farm of blade servers. The application would be re-coded so that instead of using multiple processes on a single machine it would use processes spread over the blade farm. Each blade would be one node of a vaste compute farm that jobs could be submitted to and results consolidated on a controlling server. This had wonderful scaling ability since a new blade could be added to the farm within minutes via automated build processes if the application required it, or could be repurposed to other uses just as quickly. No direct user logins were needed, and no SA work would be needed beyond replacing faulty hardware and manging what blades were assigned to what applications. To this end the SAs engineered a tightly locked down minimal access solution. It could be deployed in minutes. Hundreds of blades were purchased and installed, ready to be purposed as the customer required. Except the application developers found themselves unable to manage their application. They realised they spent a lot of time accessing the box and fiddling. They couldn't debug issues without direct access. They demanded shell access. They required additional packages. They stored unqiue state on each machine so automated builds were no longer viable. All of a sudden the SAs found themselves managing 500 additional servers rather than a blade farm and other divisions have also signed up for the service. Server population has effecitvely exploded and this is causing datacenter issues for power and heat.

[ Tom's reply: Added to limo_ch02.tex. Status: DONE ]