戴尔科技：2021年服务器趋势和观察报告（英文版）（10页）.pdf

1、Rich telemetry foraaS infrastructure components becomes even more important as the service provider needs to monitorCPU,storage,network applications,response times and other key performance metrics to ensure the customer gets theexpected performance and capabilities ofthe services to which they have

2、 subscribed. Rich telemetry will also providethe means for the service provider to remediate infrastructure issues preemptively With the emergence ofAIOpS，AImanagement ofdata and infrastructure become foundational tenets for on-premises aaSand enhanced managementhardware updates/upgrades） with minim

3、al impact to the customer workloads and operationstheme with ProjectAPEXServer Growth ls Building Vertically2Authors:Alan Brumley (Sr Distinguished Eng）， Joe Vivio (Fellow）Customers have begun embracing the scalable sofware design metodology ofusing containers and microservices aspapaau se soolas sn

4、oueA LIM ajeifalu KiyoInb o KIqIxal eu slauosn SAB pouau SIuL Kudosolrud uDrsop alos eand changes how servers are consumed.and deploying them as vertical chunks in the datacenter As the storage needs growmore servers or storage nodes areadded to increase capacity and scale performance.The storage se

5、rvice might require a protocol ofload,via an FPGA orAland ML are good examples of this freedom to innovate independently and provide services back to the datacenter.solution stacks，they allow customers to choose the size of the Al function needed in the datacenterAtasmallerscale，weexpect similardepl

6、oyment strategies attheedge，outsideof thedatacenter.Using5Gasanexample，we are seeing the architecture allow scale up and scale down in certain networking functions inside of theall hardened to deal with the environmental challenges outside of the datacenter.Each ofthese functions can beindependenty