Using Telemetry Data in Google Cloud Lift and Shift Migrations
Transcript
[Music] hey how's it going steve thousand from the cto advisor in this sponsor intel video we're outside of lake travis yes uh and austin uh as part of the cto advisor road trip we're catching up with wade holmes who's been on the program before wade is uh you're over the global solutions for the vmware solution at google correct and yes awesome to have you out here in austin keith this is my hometown here and it was exciting when i heard you were coming in and took
the opportunity to come talk to you so we're talking about telemetry data in the public cloud you've heard the conversation before in the cto advisor hybrid infrastructure intel has talked about network io [Music] cpu telemetry data and using this telemetry data in a later video we'll talk to twitter about how they're using that but we want to bring this kind of to the customer perspective everyone is looking at ways to leverage hybrid infrastructures the ability to expand their private data center and their public cloud assets
talk to me about how google tackles that problem with intel using telemetry data so [Music] every organization wants to understand their infrastructure and they frequently don't have that information it's either out of date in their existing cmdbs or they may not have information around application workloads that they may not even realize exist in their environments at times and my world's all around cloud migration and when we look at cloud migrations at google cloud somebody's telemetry data is key and we have a number of tools at
google cloud that help with gathering that data and then allows us to streamline a migration allows us to help a customer look at their performance on-premises and optimizing google cloud and allows them to save money based on this discovery data that they find in their environment on premises so one of the problems that i've run into in helping customers with migrations is that we can be a little sloppy on on premises like i can say oh i'll have eight v cpus for application that only needs
two cpus that cost me real money in public cloud and it can actually impact performance as well how do you guys help mitigate that risk because i'm using intel on premises i'm using intel in the public cloud but if i'm miss sizing my workloads that the you know you have a uh shirt on do you even lift and shift lift and shift gets a bad reputation because if i lift and shift what exactly what i have on premises those eight pc using to google cloud i'm
gonna get a bill that i'm not gonna be happy with yeah so this is where we telemetry data will allow you to lift and shift or lift shift and optimize and it's kind of a blast for the past for me um a lot of people may know they have a background with vmware and back in the old days when environments were going from physical servers to virtual servers this was the same exercise that was required vmware used to have a tool capacity planner they would come
in and gather this data they would find oh you have all these physical machines that are only 20 utilized so now we have a more optimized environment that may be virtualized on premises but as mentioned it's still not unnecessarily right size to be most efficient so we have a tool at google cloud called stratazone google cloud stratazone which can gather data from workloads at the os level and or gather data from vcenter servers for vmware environments and look at all the usage useful to all the
spikes and peaks from a cpu utilization perspective memory perspective storage can actually fingerprint applications and specific databases now you have all this understanding of your environment and then you can now as you lift and shift also have the option to optimize those workloads as you move them into cloud so wait talk to me about the importance because a lot of times we just look at cpu and memory as the two main knobs that we have in the data center but io you just mentioned you know
uh uh traffic in between workloads etc how important is it to understand the i o in my current data center before i lift and shift or lift and improve workloads in the public cloud and how are you working with intel to surface up that data yeah so it's absolutely critical to mention there may be different storage architectures they may have fiber channel attached storage they may have iscsi there may be different dependencies from a storage architecture that is going to now have to be translated into
how you gain that same performance in cloud or it may be groups of applications that need to communicate with each other because they have some need for locality to have extremely low latency within their existing data center posture you need to understand all of these things and as you gather data on premises you can are the tooling that we have today that can gather all this either i said os level and it uses either if it's windows os it goes in and gathers things by wmi
or if it's linux os it goes in and gathered by ssh you are able to now gather that specific information at the os and application layer and understand things from a top down perspective that includes that storage data and includes that network and data and how the applications actually communicate now i'm talking about actualizing this uh you right before we started recording you talked about how google is working with partners to help this real make this real for customers so google and intel have partnered together
to be able to have partners accelerate our customers workloads we have a program called 1k 90 and 10k 90 that allows for partners to accelerate your migration in 90 days so this is something that if you reach out to um you're working with google you can find out more from your google account reps and this is something that allows for your migration to be more seamless and efficient and all our partners are enabled to be able to apply this program and i'm a former vmware uh
fan not a former currently i still love vmware where can people find out more about the vmware solution running on google's solution oh so cloud.google.com look up google cloud vmware engine and there's a plethora of information out there and or you can ping me on twitter at w homes if you have a specific question and i'm always happy to reach out well thank you again for watching this series focusing on telemetry data in the enterprise data center and how we're using that to manage the hybrid
infrastructure specifically in the cto advisor infrastructure you want to find out more about the cto advisor you can follow me on the web the ctoadvisor.com is the website and you can hit me up on twitter dms are open i'm brave at ctoadvisor on twitter talk to you next cto does
