Interesting article in Information Week yesterday – What’s killing APM
Those experiencing outages on a daily basis went from 8% in 2010 to 10% now, and those experiencing monthly outages went from 24% in 2010 to 27% now.
>>Wow….has this become like the blue screen of death? We’re being educated by the application providers to expect outages in the apps that we use? I don’t accept that. I just think that application infrastructures have become so complex that IT is having a hard time managing them appropriately. They need help!
As app environments become more dynamic and lifecycles for apps shorten, the substantial effort required for APM isn’t worth the already iffy results it provides.
One area where APM tools definitely have let IT pros down is in keeping up with complexity. As apps take a more service-oriented design, the task of setting up an APM tool to give anything close to meaningful information is much harder than it used to be. As one survey respondent put it: “There is a major conundrum related to the real-world use of APM tools. They work quite well in a static infrastructure environment. Unfortunately, current APM tools do not work well in dynamic Web services, and public and private cloud-based infrastructures, since they depend on statically defined relationships. By the time these relationships are defined to the APM tool, they will in all probability be obsolete, thereby negating the value and relevance of the APM tool.”
>>I agree. The vendor community needs to rethink the APM model. Here are some of the issues that traditional APM approaches struggle with:
1. Modern apps have dynamic, indeterminate infrastructures
2. Modern apps are not single language, they are increasingly Polyglot
3. All layers of the “app stack” cannot be instrumented as in the past (think about the network layer in cloud as one simple example)
4. Apps run across multiple providers….time-stamping is often a problem (different providers have different clocks)
5. Application code changes very frequently; multiple times every day in many cases
This is exactly the problem that we are trying to solve at Boundary. Create something so simple to deploy and cost effective that it’s a no-brainer to implement. Something that is real-time (not polling or sampling). Something that does not require a “defined application topology” to operate but instead has a constantly changing topology as the application itself changes. Something that works across Polyglot apps….not just Java or .Net or Ruby (Erlang anyone?)
Some of the differences that we see between traditional and next gen APM….
And finally, nobody should be surprised. Management tools always lag the application infrastructure by several years which creates customer frustration (it’s just a fact of life though….bought my daughter an iPhone5 but cannot find a case for it – same thing).
And, the big vendors are not the ones that step up. They have much slower product cycles and are not in-tune with early adopter customers. Innovation in management will always come from small vendors.
Look up your neighbor (if you care), your employer, your zip code.
Innovative use of product and nice marketing
CA Inc. (CA) on Monday sued a software company founded by one of its former employees for allegedly infringing three of its patents.
The Islandia, N.Y. company accused privately held New Relic Inc. of reusing the some of the intellectual property CA acquired through its 2006 purchase of Wily Technology. Wily founder Lew Cirne joined CA through the acquisition and remained an employee until 2007. He started New Relic in San Francisco in 2008…..
Sort of thinking about the whole theme of Cloud based IT services and in fact more than that hybrid application infrastructures. It’s not just about the Cloud, it’s about all the other many pieces that come together to make up today’s applications.
Are any MSPs offering monitoring services that can deal with this? Something that can give you early warnings of EC2 problems and the affect they have on your app, understanding when 3rd party APIs are running slow, combining that with the on-premise part of the application and rolling it all together to understand end user experience. And of course, being able to swiftly and painlessly migrate/move part of your infrastructure when problems do occur?
Is anyone doing this or is everyone too busy offering remote monitoring for MS Exchange etc? Love to hear from you….email me
Joe P(@mspmentor)….Charles and Celia W(@mspalliance)…..know of anyone pushing into modern application infrastructures?
If not, then let’s start one!
Watch the Boundary blog today, we’ll be writing up another Cloud outage that was spotted by a Boundary customer many hours before the Cloud provider sent notification.
My point here is NOT to bash the Cloud providers. Without being sure but I would think they have better availability and performance of almost any data center worldwide – Cloud is the future and the providers are providing incredible benefits.
But instead it is to educate folks that you cannot absolve yourself of the responsibility for ensuring that your customers are getting great performance. That is not the responsibility of the Cloud provider; it is your responsibility. Customer satisfaction is always your responsibility and just because you are using Amazon, Microsoft, Google, Rackspace, Softlayer, EngineYard or anyone else – you still have the responsibility.
Two pieces of interesting news today.
First, Riverbed announced that they are acquiring OPNET for close to $1bn. We congratulate both companies on this move; it seemed as though OPNET’s business growth had slowed to about 6% license revenue growth in recent times (appears to be slower than the overall market?), but I’m sure Riverbed will be hoping to kick that up again as a result of bundling with their other products.
It’s a clear indication to everyone, that the sourcing and analyzing of the “flow” data is of strategic importance for many different use cases.
What was also interesting was the following article http://www.forbes.com/sites/joemckendrick/2012/10/28/two-thirds-of-the-worlds-data-centers-going-to-cloud-cisco – headlined as “two thirds of the world’s datacenters are going to Cloud : Cisco” – and at the same time global datacenter traffic will grow fourfold over the next four years.
All of this appears to be good news for Boundary’s strategy.
We believe that we are the only company that is able to collect and analyze the flow data from hybrid IT environments (cloud) as though they were a single homogenous environment; we think that the dataset that Boundary is collecting and exposing is the greatest machine generated dataset that has yet to be fully explored and eventually will surpass all other datasets for determining application performance and security.
Riverbed shows us a billion reasons why they believe that data is important and Cisco is telling us that by 2014 the majority of the workloads will be in the cloud.
I’ve been on this weight-loss kick for a little while, being somewhat successful but still a long way to go. My target is 200 lbs or thereabouts by thxgiving. Lost about 12-14 so far and another 18 to go.
Anyhow, decided this morning to see if I could run on the treadmill for the entire Chelsea-Man Utd game. I started about 10 mins ahead of time and I ran (with occasional walking) for 2 hours.
I actually did 10.5 miles – I was wanting to keep going at the end (but I had to go and pick my son up) so that I could get to 13 miles – I figured that would be cool to do the equivalent of a half marathon.
Now, anyone that knows me knows that (a) I’m overweight, currently around 218 lbs (15 1/2 stone) (b) I don’t run very often (c) I have done absolutely no training.
Must admit my legs are aching like anything the rest of the day (advil!) and I am not looking forward to tomorrow morning but I’m happy. Really wish I had had the time available to keep going….reckon I’d have done my 1/2 marathon in about 2 1/2 hours.
Next weekend – I’m going to go for 13 miles. The key for me is to not get bored….watching football enables me to do so.
In my inbox this morning. The emails started:
1. CA reported disappointing F2Q13 results as both the macro environment and factors specific to CA contributed to the underperformance for the second consecutive period.
2. SWI (SolarWinds) easily exceeded 3Q expectations
Just one data point among many of course, but it’s pretty clear to me.
New wave IT mgmt vendors are growing like crazy: SolarWinds, Service-Now, Splunk, New Relic, AppDynamics. Incumbents are actually shrinking.
The shift to new IT infrastructures is causing the incumbents to be rendered legacy. They need to acquire and then get out of the way and let the new vendors run/grow their own business.
Two problems though:
First, the traditional methods that the incumbents use to value companies for acquisitions need to go out the window. They have to realize that they are in a death match and fighting for future relevance. How many spreadsheets did Facebook run when it acquired Instagram (OK, maybe an extreme example but you get my point) – maybe VMWare/Nicira is a better example.
Second, for some reason, they find it really difficult to leave the acquired companies alone.
e.g. take a read of this article published today http://www.cxotoday.com/story/cultural-integration-is-key-to-our-successful-acquisitions/ and in particular this paragraph….
Acquisitions bring along challenges such as culture clash. How has CA been able to address these issues inspite of its acquisition spree?
This is something CA has continued to better over the years. I myself joined CA, when it had acquired Netegrity. What CA has learned to do is to provide a little autonomy to the acquired organizations and give them time to get used to CA’s culture. For example, couple of years ago when we acquired Nimsoft, we ran it as an autonomous unit and only now have we merged it with proper CA. During the last two years, we worked together and closer with the Nimsoft team and also enabled Nimsoft to leverage the assets of CA and the relationship with CA without losing the integrity of Nimsoft’s spirit.
Another reason we are successful at acquisitions is because we integrate the culture of the new organizations. However, this is not easy and neither it works all the time. We also lose a few individuals in the process but we still manage to retain a major chunk of our human resource.
Uh oh, pinocchio.
The shift is on….
Blog post from Molly last night….http://boundary.com/blog/2012/10/23/boundarys-observations-during-the-amazon-degradation-yesterday/
I simply do not understand anyone that hosts business critical applications in any cloud and does not invest the time and relatively small $$ to monitor the service that is being delivered.
Yeah I know we all read “The Big Switch” and we all have this vision that computing is as ubiquitous as electricity but guess what…..it is nowhere near as reliable.
This is not an Amazon problem; I would hazard a guess that they are probably one of the best run compute environments on the planet. But, it’s a fundamental reality of IT that is not going away any time soon.
IT is advancing at a rapid pace and will continue to do. Therefore we should expect that in an environment of constant change and innovation, reliability will not be 100%.
Now, you could say that Amazon could and should have posted details of the outage a little sooner than 2 hours after it occurred (are you kidding me….2 hours is a lifetime for a business that depends on AWS), and could and should have resolved the problem much sooner (I’m sure they did their best) but, the reality of the situation is….
Why would you risk your business on the ability of someone else to deliver and not even monitor their performance?
EVERY company that has business critical applications in ANY cloud provider needs their own monitoring. Otherwise…well, caveat emptor.
Shameless plug…Boundary has a free offering that makes it really easy for most EC2 customers to monitor their applications for no cost. So, there is no excuse.