cancel
Showing results for 
Search instead for 
Did you mean: 
Create Post

It’s the cloud, so problem solved!

Level 11

In the last two posts, I talked about databases and why network and storage are so darn important for their well-being. These are two things that a lot of companies are struggling with and in some ways, seem to have mastered over the years in their own data centers. But now it seems they will be obsolete when things like cloud are quickly becoming the new normal.

Let’s leave the automotive comparison for now and let’s concentrate on how cloud strategy is affecting the way we use and support our databases.

There are a lot of ways to describe the cloud, but I think the best way for me to describe it is captured in the picture below:

CmntBpt.png

As Master Yoda explains, cloud is not something magical. You cannot solve all of your problems by simply moving all your resources there. It will not mean that once your database is in the cloud that your storage and network challenges will magically vanish. It is literally just another computer that, in many cases, you will even share with other companies that have moved their resources to the cloud as well.

As we all know, knowledge is power. The cloud doesn’t change this. If we want to migrate database workloads to the cloud, we need to measure and monitor. We need to know the impact of moving things to the cloud. As with the Force, the cloud is always there ready to be used. It is the how and when that will impact the outcome the most. There are multiple ways to leverage the cloud as a resource to gain an advantage, but that doesn’t mean that moving to the cloud is the best answer for you. Offense can be the best defense, but defense can also be the best offense. Make sure to know thy enemy, so you won’t be surprised.

May the cloud be with you!

37 Comments

We've moved a lot of data to the cloud. O365, Exchange and so many Outlook accounts & files.  It's not a fast and reliable solution for us yet, and I wish we hadn't gone there.  But the carrot dangled in front of us ($6M in MS license costs saved) outweighed any lack of knowledge of poor performance and reliability/availability.  Here's hoping MS becomes better, faster, more available . . .  But in this day of DDOS and Ransomware and the like, my organization is already preparing simulations where we have to shut down the Internet pipe and isolate our data centers to stop a hypothetical Ransomeware spread.

I believe most of our 17,000 employees don't understand our reliance on the cloud, now that we've moved so much functionality there.  They trust management to make the right decisions for them at the highest level, and budget is a powerful incentive to move.

The Yoda I remember had more than one comment about a cloud:

pastedImage_0.png

Level 9

Don't forget "Patience you must have..  My young Padawan."  when it comes to utilizing the cloud.  

Sigh.  When the users call to complain, when e-mail won't open, when O365 doesn't work . . .

Yep.  "Patience."   It's what we need when we'd rather get our jobs done instead of waiting for the cloud to deliver our apps & files.

Thanks goodness SW apps show the problems not in MY network!

Level 15

gramelins .png  

Level 14

I was told a long time ago that "All computers wait at the same speed" and it makes no difference it it's your platform, someone else's, dedicated or shared.

Bad code, bad database design or bad application design do not just go away because you think that you've abdicated responsibility to someone else.

You need to monitor it before & after a migration. Perhaps it will be more accessible, but will it work as well as it did in-house? It can be more scale-able, but will it actually be cheaper?

Remember that there's no such thing as a free lunch.

TANSTAAFL!  A great word from The Moon Is A Hard Mistress.

No matter how fast any computer waits in comparison to another computer, tit's nice that SW tools show how some systems may present slower user experiences due to problems with database or spinning disc or interface errors or CPU/RAM or poorly written code.  Even NetPath proves that performance varies depending on ISP and WAN carriers.

Setting thresholds and keeping monitoring going, and using Alerts to let us know when something has changed and impacted the end users' experience is a great way to justify our tools and our value to the company.

Level 14

Yep, been banging on about this to various IT managers for years.  "The Cloud" isn't the answer to your problems.  It just reduces / removes the need for hardware refresh and lowers the requirement for on-site power and air-con.  It does however increase the need for good systems monitoring and vendor management.  It also means potentially longer downtime in the event of a problem as you no longer control all aspects of your IT.  Someone else, who has no real interest in your systems, now controls the hardware and the access.  Management also forget the need for a bigger internet pipe and then complain when things are slower and outages have more impact despite the fact that they forced the move against their experts express wishes and warnings.  I used to have a full head of hair.

MVP
MVP

What are you monitoring for Office 365? We've moved our email there as well.

MVP
MVP

Nice write up

MVP
MVP

My favorite perfect answer to shoot down.

We moved our email to O365 - the funny part is that our users call with issues, expecting us to be able to fix things the way we used to.

Now our answer is that we are "saving money" by using the cloud which means all those neat little workarounds that support used to do are no longer possible.

Short answer, yes sir/madam this is now expected behavior. This is how long it takes.

Works great for me, I don't have to manage exchange

Level 14

We have 50,000 accounts in O365 (all students).  In the process of moving the 5000 staff accounts from on prem exchange to O365.  They won't be happy with the loss of functionality, slower performance and IT's newfound inability to provide the same level of service.  Management are happy so that's OK then.  Ho hum.

Level 10

Interesting point.

In my opinion, the core of the problem stays the same. We are using the same platform (let it be DB, Web, App, Network), but only building more and in a much faster pace. If one needs to optimize or troubleshoot or migrate these resources, one would still have to take the same approach but just have to use different tools.

Level 20

I do have to admit that between exchange and now the integration with skype it can be difficult totally switching the entire business over to skype for calls across the entire MPLS network.  The QOS alone is making us look at some new end to end performance management products with a focus on VOIP and skype.

Level 14

I'd be interested on O365 monitoring as well.. Just migrated 3 weeks ago

Level 14

Interested in what you come up with...About to deploy RingCentral and Skype.

When we started using the cloud I set up NetPath to watch the target/destinations shared with me.  They include:

  • powerpoint.officeapps.live.com
  • view.officeapps.live.com
  • visio.officeapps.live.com
  • word-edit.officeapps.live.com
  • word-view.officeapps.live.com
  • officeapps.live.com
  • broadcast.officeapps.live.com
  • onenote.officeapps.live.com
  • view.officeapps.live.com

The obvious drawback for me, when relying on NetPath monitoring, is when a cloud service changes its DNS entry, adds another one, or removes one or more.

Here's a typical example of NetPath showing current and historical problems with a cloud service--this one happens to be powerpoint.officeapps.live.com:

pastedImage_16.png

In my experience, Management is initially happy if they can save significant money for licenses locally versus sending O365 or Outlook / Exchange to the cloud.

However, in my organization, we're seeing so many complaints, and have only migrated about 5,000 staff members (out of 17,000 planned to move to O365), that Management is very concerned about Cloud performance. 

I have a strong suspicion there'll be funding for Solarwinds four new cloud monitoring applications as a result.  This in spite of my recommending NOT moving to the cloud without being able to verify (monitor) to ensure we can trust their claims of reliability, availability, security, and resilience.

MVP
MVP

Awesome, Thanks I'm setting them up now.

I'd love to hear what you see on them after a day or a week.  Please IM me, or post your findings in this thread / forum. 

Level 14

These are awesome... Going to try one or two today......Thanks rschroeder

And has your organization moved to Solarwinds cloud monitoring solutions?

Level 20

One of the neatest new tools I've seen for netflow is liveaction.  The product was originally developed for the Navy to allow them to visualize all the netflow, IPFIX, nbar, etc. in their flows so they could reconfigure each ship as it goes into different ports all around the world and adapt to each unique network situation.  I have it running on a few of our sites now in a POC.  I still am keeping NTA but this will help me with my QOS and it also does the netflow configuration for you if you want and lets you access access-lists on the fly from the GUI.  I've been pretty blown away by it.  It's already helped me find places were QOS wasn't working the way it should be.  Also since now I may have two, three or maybe more tools to analyze flows I'm using a flow forwarder called:  GitHub - sleinen/samplicator: Send copies of (UDP) datagrams to multiple receivers, with optional sa...

Level 20

We're investigating the GCC High Office365 offering.  At least they have everyone background checked and are all US Citizens:

GCC High and DoD

It will only be for controlled unclassifed information but may be promising... idk after hearing from everyone here... hehe!

This sounds like a great Feature Request.  If you create it, I'll vote it up!

Level 21

We use SharePoint as part of 0365 and I have a NetPath running to that and it's pretty clean, every now and again I will see a red (no data) poll but I honestly haven't been able to verify if that is a problem on my end or somewhere else.  With that being said we also have direct peering with the Six up in Seattle so we have incredibly good connectivity with Microsoft.

Level 21

We used Pingdom years ago but I haven't really used any of these tools recently.  I would love to but I haven't yet had a good use case for them.

Level 21

ecklerwr1​ do you know if there are any good reference architectures out there for a good Skype implementation as a phone replacement solution?

"Direct peer with the Six up in Seattle?"

pastedImage_0.png

I'm still on the fence, in many instances I've seen it merely change the pain point.

Level 21

<shameless sales plug> We have 3 data centers we operate and all of them have colocation space available and we also have local private cloud based hosting as well, both options have this same level of connectivity.  </shameless sales plug>

Level 14

thank you sir!

Level 13

Cool stuff. 

As the Cloud providers come parading in to our office promising near-unlimited computing capability to our executives I have witnessed a fundamental shift in mindset. Prior to the Cloud capacity constraints was the sole responsibility of CIO's and infrastructure teams. Now with the Cloud it becomes the responsibility of the providers and meeting their SLA's. The CIO has a scapegoat when the provider doesn't deliver.

Level 14

It's all so nieve. You can outsource it, you can cloud it, you can use an SLA to blame someone else, but it's YOUR business that suffering, it's your job that's at risk and it will always stay as being YOUR responsibility. I really don't understand why the c-list folk can't see that when it's so obvious to us simple techies.

& that's why I use Pingdom, WPM & other modules to keep them honest. That's both the supplier & our c-list. 🙂

Level 17

Nice write up, and all to true. Trust but Verify comes to mind; or Trust after Verifying.

Level 13

And the issues are often brand new, but of course no less complex. Couldn't agree more.

very nice demonstration of real life Netpath usage! Thanks for this

About the Author
In the IT since 1998 and enjoying every last bit of it. The last few years are mainly focused on virtualization and Storage. VMware VCAP-DCA, VCP 4/5, VSP 4/5, VTSP 4/5, MCSA, MCTS, MCP, CCA and CCNA