Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

classic Classic list List threaded Threaded
12 messages Options
Reply | Threaded
Open this post in threaded view
|

Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

AlexR

We have a really good pilot with ES and now infrastructure people pushing google search on us because they have in on their approved stack (and I bet to charge for licenses too)

What are the weaknesses of google search compared to ES?

Our application works with heavily structured data (large number of small fields) and has lots of faceting and analytics. Is google search any good with it structured documents with many fields? Does it have it advanced faceting? Does it allow fine grained control over indexing?


--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

Tinou Bao
I haven't used Google search. Briefly played around w/ Amazon CloudSearch. One thing is latency that is probably applicable to any cloud based solution. Within a data center you can update ES and have near realtime updates. With AWS CloudSearch I remember it does things in batches and there's also the latency of a network request to Virginia.

On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:

We have a really good pilot with ES and now infrastructure people pushing google search on us because they have in on their approved stack (and I bet to charge for licenses too)

What are the weaknesses of google search compared to ES?

Our application works with heavily structured data (large number of small fields) and has lots of faceting and analytics. Is google search any good with it structured documents with many fields? Does it have it advanced faceting? Does it allow fine grained control over indexing?


--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

AlexR
Thanks Tinou,

They have (or would have) google search (I guess appliance) in the data center. so i do not think latency will be an issue. it is the services it can provide, and integration I need to understand.
From very briefly going over google documentation my feeling is:

1.       It is not free like ElasticSearch. The licenses may be quite costly

2.       It is optimized for document and site indexing and searching and not so much for structured data indexing searching, analytics and faceting while elastic allows very easy mixing of both

3.       It is a shrink wrap product which comes with GUI, administration and is centered around searching for documents or web pages or emails plus some metadata about them not structured data with 100th of fields with need to

4.       It is not an open source or open platform - it is a shrink wrapped solution with various connectors to pull data from various data sources.

a.       It may be impossible to do something unless google baked it. For example showing which fields got hits might not even be possible

b.      It will be harder to integrate

5.       Our product will be constrained to what google platform offers and we will be bound to versions, licenses and upgrade schedule the IT Shop has for the product

6.       All documentation I found is about configuring and managing nothing much about developing for it, executing complex custom queries etc.

7.       I checked its user groups – they seems to be dormant unlike elastic search. Not sure where one can get any support unless it is paid

8.       Last not the least it will be starting development all over even if it is possible to achieve the same with google 


On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <[hidden email]> wrote:
I haven't used Google search. Briefly played around w/ Amazon CloudSearch. One thing is latency that is probably applicable to any cloud based solution. Within a data center you can update ES and have near realtime updates. With AWS CloudSearch I remember it does things in batches and there's also the latency of a network request to Virginia.


On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:

We have a really good pilot with ES and now infrastructure people pushing google search on us because they have in on their approved stack (and I bet to charge for licenses too)

What are the weaknesses of google search compared to ES?

Our application works with heavily structured data (large number of small fields) and has lots of faceting and analytics. Is google search any good with it structured documents with many fields? Does it have it advanced faceting? Does it allow fine grained control over indexing?


--
You received this message because you are subscribed to a topic in the Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
To unsubscribe from this group and all its topics, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

Jaguar

Our infra guy also provide Google search service. One issue, if not miss configuration,is that the search result is not stable. You could find the result one day, but failed the other day.

My 2 cents here.

在 2013-5-24 上午7:10,"Alex Roytman" <[hidden email]>写道:
Thanks Tinou,

They have (or would have) google search (I guess appliance) in the data center. so i do not think latency will be an issue. it is the services it can provide, and integration I need to understand.
From very briefly going over google documentation my feeling is:

1.       It is not free like ElasticSearch. The licenses may be quite costly

2.       It is optimized for document and site indexing and searching and not so much for structured data indexing searching, analytics and faceting while elastic allows very easy mixing of both

3.       It is a shrink wrap product which comes with GUI, administration and is centered around searching for documents or web pages or emails plus some metadata about them not structured data with 100th of fields with need to

4.       It is not an open source or open platform - it is a shrink wrapped solution with various connectors to pull data from various data sources.

a.       It may be impossible to do something unless google baked it. For example showing which fields got hits might not even be possible

b.      It will be harder to integrate

5.       Our product will be constrained to what google platform offers and we will be bound to versions, licenses and upgrade schedule the IT Shop has for the product

6.       All documentation I found is about configuring and managing nothing much about developing for it, executing complex custom queries etc.

7.       I checked its user groups – they seems to be dormant unlike elastic search. Not sure where one can get any support unless it is paid

8.       Last not the least it will be starting development all over even if it is possible to achieve the same with google 


On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <[hidden email]> wrote:
I haven't used Google search. Briefly played around w/ Amazon CloudSearch. One thing is latency that is probably applicable to any cloud based solution. Within a data center you can update ES and have near realtime updates. With AWS CloudSearch I remember it does things in batches and there's also the latency of a network request to Virginia.


On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:

We have a really good pilot with ES and now infrastructure people pushing google search on us because they have in on their approved stack (and I bet to charge for licenses too)

What are the weaknesses of google search compared to ES?

Our application works with heavily structured data (large number of small fields) and has lots of faceting and analytics. Is google search any good with it structured documents with many fields? Does it have it advanced faceting? Does it allow fine grained control over indexing?


--
You received this message because you are subscribed to a topic in the Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
To unsubscribe from this group and all its topics, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

louis chabardes

To my knowledge, GSA doesn't have any faceting/analytics/counts capabilities: http://www.google.com/support/enterprise/static/gsa/docs/admin/70/gsa_doc_set/xml_reference/



On Thu, May 23, 2013 at 6:03 PM, xiong.jaguar <[hidden email]> wrote:

Our infra guy also provide Google search service. One issue, if not miss configuration,is that the search result is not stable. You could find the result one day, but failed the other day.

My 2 cents here.

在 2013-5-24 上午7:10,"Alex Roytman" <[hidden email]>写道:

Thanks Tinou,

They have (or would have) google search (I guess appliance) in the data center. so i do not think latency will be an issue. it is the services it can provide, and integration I need to understand.
From very briefly going over google documentation my feeling is:

1.       It is not free like ElasticSearch. The licenses may be quite costly

2.       It is optimized for document and site indexing and searching and not so much for structured data indexing searching, analytics and faceting while elastic allows very easy mixing of both

3.       It is a shrink wrap product which comes with GUI, administration and is centered around searching for documents or web pages or emails plus some metadata about them not structured data with 100th of fields with need to

4.       It is not an open source or open platform - it is a shrink wrapped solution with various connectors to pull data from various data sources.

a.       It may be impossible to do something unless google baked it. For example showing which fields got hits might not even be possible

b.      It will be harder to integrate

5.       Our product will be constrained to what google platform offers and we will be bound to versions, licenses and upgrade schedule the IT Shop has for the product

6.       All documentation I found is about configuring and managing nothing much about developing for it, executing complex custom queries etc.

7.       I checked its user groups – they seems to be dormant unlike elastic search. Not sure where one can get any support unless it is paid

8.       Last not the least it will be starting development all over even if it is possible to achieve the same with google 


On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <[hidden email]> wrote:
I haven't used Google search. Briefly played around w/ Amazon CloudSearch. One thing is latency that is probably applicable to any cloud based solution. Within a data center you can update ES and have near realtime updates. With AWS CloudSearch I remember it does things in batches and there's also the latency of a network request to Virginia.


On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:

We have a really good pilot with ES and now infrastructure people pushing google search on us because they have in on their approved stack (and I bet to charge for licenses too)

What are the weaknesses of google search compared to ES?

Our application works with heavily structured data (large number of small fields) and has lots of faceting and analytics. Is google search any good with it structured documents with many fields? Does it have it advanced faceting? Does it allow fine grained control over indexing?


--
You received this message because you are subscribed to a topic in the Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
To unsubscribe from this group and all its topics, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

joergprante@gmail.com
In reply to this post by AlexR
To add one more important aspect: performance. It should be possible to
order an evaluation license or something for a limited time (1-3
months). Put your data on the product (if possible) and measure the
"documents per second" while indexing and "queries per second" while
searching, for all your workloads. Compare that numbers to ES on your
favorite platform. Add your findings (ease of administration, network
latency, resource sizing challenges etc) Present the result to
management. Management always listens to technical staff when
performance is the issue - it is expensive to fix performance issues later.

Another more important one to kill every cloud-based app scenario is
data privacy but I'm not sure if your data is relevant to this topic.

Jörg

Am 24.05.13 01:10, schrieb Alex Roytman:

> Thanks Tinou,
>
> They have (or would have) google search (I guess appliance) in the
> data center. so i do not think latency will be an issue. it is the
> services it can provide, and integration I need to understand.
> From very briefly going over google documentation my feeling is:
>
> 1.It is not free like ElasticSearch. The licenses may be quite costly
>
> 2.It is optimized for document and site indexing and searching and not
> so much for structured data indexing searching, analytics and faceting
> while elastic allows very easy mixing of both
>
> 3.It is a shrink wrap product which comes with GUI, administration and
> is centered around searching for documents or web pages or emails plus
> some metadata about them not structured data with 100^th of fields
> with need to
>
> 4.It is not an open source or open platform - it is a shrink wrapped
> solution with various connectors to pull data from various data sources.
>
> a.It may be impossible to do something unless google baked it. For
> example showing which fields got hits might not even be possible
>
> b.It will be harder to integrate
>
> 5.Our product will be constrained to what google platform offers and
> we will be bound to versions, licenses and upgrade schedule the IT
> Shop has for the product
>
> 6.All documentation I found is about configuring and managing nothing
> much about developing for it, executing complex custom queries etc.
>
> 7.I checked its user groups – they seems to be dormant unlike elastic
> search. Not sure where one can get any support unless it is paid
>
> 8.Last not the least it will be starting development all over even if
> it is possible to achieve the same with google
>
>
> On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <[hidden email]
> <mailto:[hidden email]>> wrote:
>
>     I haven't used Google search. Briefly played around w/ Amazon
>     CloudSearch. One thing is latency that is probably applicable to
>     any cloud based solution. Within a data center you can update ES
>     and have near realtime updates. With AWS CloudSearch I remember it
>     does things in batches and there's also the latency of a network
>     request to Virginia.
>
>
>     On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:
>
>
>         We have a really good pilot with ES and now infrastructure
>         people pushing google search on us because they have in on
>         their approved stack (and I bet to charge for licenses too)
>
>         What are the weaknesses of google search compared to ES?
>
>         Our application works with heavily structured data (large
>         number of small fields) and has lots of faceting and
>         analytics. Is google search any good with it structured
>         documents with many fields? Does it have it advanced faceting?
>         Does it allow fine grained control over indexing?
>
>
>     --
>     You received this message because you are subscribed to a topic in
>     the Google Groups "elasticsearch" group.
>     To unsubscribe from this topic, visit
>     https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
>     To unsubscribe from this group and all its topics, send an email
>     to [hidden email]
>     <mailto:elasticsearch%[hidden email]>.
>     For more options, visit https://groups.google.com/groups/opt_out.
>
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "elasticsearch" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to [hidden email].
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.


Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

AlexR
In reply to this post by louis chabardes
Thanks Louis, looks like you are correct. I can't find it. I thought it was dynamic results clustering but I do not think it is. I read some articles where people were talking about faceting in GSA but i do not see it in the docs 

On Thursday, May 23, 2013 9:26:10 PM UTC-4, Louis Chabardes wrote:

To my knowledge, GSA doesn't have any faceting/analytics/counts capabilities: http://www.google.com/support/enterprise/static/gsa/docs/admin/70/gsa_doc_set/xml_reference/



On Thu, May 23, 2013 at 6:03 PM, xiong.jaguar <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="ZWe9fmTjch8J">xiong....@...> wrote:

Our infra guy also provide Google search service. One issue, if not miss configuration,is that the search result is not stable. You could find the result one day, but failed the other day.

My 2 cents here.

在 2013-5-24 上午7:10,"Alex Roytman" <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="ZWe9fmTjch8J">royt...@...>写道:

Thanks Tinou,

They have (or would have) google search (I guess appliance) in the data center. so i do not think latency will be an issue. it is the services it can provide, and integration I need to understand.
From very briefly going over google documentation my feeling is:

1.       It is not free like ElasticSearch. The licenses may be quite costly

2.       It is optimized for document and site indexing and searching and not so much for structured data indexing searching, analytics and faceting while elastic allows very easy mixing of both

3.       It is a shrink wrap product which comes with GUI, administration and is centered around searching for documents or web pages or emails plus some metadata about them not structured data with 100th of fields with need to

4.       It is not an open source or open platform - it is a shrink wrapped solution with various connectors to pull data from various data sources.

a.       It may be impossible to do something unless google baked it. For example showing which fields got hits might not even be possible

b.      It will be harder to integrate

5.       Our product will be constrained to what google platform offers and we will be bound to versions, licenses and upgrade schedule the IT Shop has for the product

6.       All documentation I found is about configuring and managing nothing much about developing for it, executing complex custom queries etc.

7.       I checked its user groups – they seems to be dormant unlike elastic search. Not sure where one can get any support unless it is paid

8.       Last not the least it will be starting development all over even if it is possible to achieve the same with google 


On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="ZWe9fmTjch8J">tino...@...> wrote:
I haven't used Google search. Briefly played around w/ Amazon CloudSearch. One thing is latency that is probably applicable to any cloud based solution. Within a data center you can update ES and have near realtime updates. With AWS CloudSearch I remember it does things in batches and there's also the latency of a network request to Virginia.


On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:

We have a really good pilot with ES and now infrastructure people pushing google search on us because they have in on their approved stack (and I bet to charge for licenses too)

What are the weaknesses of google search compared to ES?

Our application works with heavily structured data (large number of small fields) and has lots of faceting and analytics. Is google search any good with it structured documents with many fields? Does it have it advanced faceting? Does it allow fine grained control over indexing?


--
You received this message because you are subscribed to a topic in the Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
To unsubscribe from this group and all its topics, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="ZWe9fmTjch8J">elasticsearc...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="ZWe9fmTjch8J">elasticsearc...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="ZWe9fmTjch8J">elasticsearc...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

AlexR
In reply to this post by joergprante@gmail.com
Thanks Jorg. If it gets that far (hope not) we will do performance trails. I do not think it will have any "cloud" issues because they will run it in the data center...

On Friday, May 24, 2013 5:42:31 AM UTC-4, Jörg Prante wrote:
To add one more important aspect: performance. It should be possible to
order an evaluation license or something for a limited time (1-3
months). Put your data on the product (if possible) and measure the
"documents per second" while indexing and "queries per second" while
searching, for all your workloads. Compare that numbers to ES on your
favorite platform. Add your findings (ease of administration, network
latency, resource sizing challenges etc) Present the result to
management. Management always listens to technical staff when
performance is the issue - it is expensive to fix performance issues later.

Another more important one to kill every cloud-based app scenario is
data privacy but I'm not sure if your data is relevant to this topic.

Jörg

Am 24.05.13 01:10, schrieb Alex Roytman:

> Thanks Tinou,
>
> They have (or would have) google search (I guess appliance) in the
> data center. so i do not think latency will be an issue. it is the
> services it can provide, and integration I need to understand.
> From very briefly going over google documentation my feeling is:
>
> 1.It is not free like ElasticSearch. The licenses may be quite costly
>
> 2.It is optimized for document and site indexing and searching and not
> so much for structured data indexing searching, analytics and faceting
> while elastic allows very easy mixing of both
>
> 3.It is a shrink wrap product which comes with GUI, administration and
> is centered around searching for documents or web pages or emails plus
> some metadata about them not structured data with 100^th of fields
> with need to
>
> 4.It is not an open source or open platform - it is a shrink wrapped
> solution with various connectors to pull data from various data sources.
>
> a.It may be impossible to do something unless google baked it. For
> example showing which fields got hits might not even be possible
>
> b.It will be harder to integrate
>
> 5.Our product will be constrained to what google platform offers and
> we will be bound to versions, licenses and upgrade schedule the IT
> Shop has for the product
>
> 6.All documentation I found is about configuring and managing nothing
> much about developing for it, executing complex custom queries etc.
>
> 7.I checked its user groups – they seems to be dormant unlike elastic
> search. Not sure where one can get any support unless it is paid
>
> 8.Last not the least it will be starting development all over even if
> it is possible to achieve the same with google
>
>
> On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <<a href="javascript:" target="_blank" gdf-obfuscated-mailto="VZCfELGD8noJ">tino...@...
> <mailto:<a href="javascript:" target="_blank" gdf-obfuscated-mailto="VZCfELGD8noJ">tino...@...>> wrote:
>
>     I haven't used Google search. Briefly played around w/ Amazon
>     CloudSearch. One thing is latency that is probably applicable to
>     any cloud based solution. Within a data center you can update ES
>     and have near realtime updates. With AWS CloudSearch I remember it
>     does things in batches and there's also the latency of a network
>     request to Virginia.
>
>
>     On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:
>
>
>         We have a really good pilot with ES and now infrastructure
>         people pushing google search on us because they have in on
>         their approved stack (and I bet to charge for licenses too)
>
>         What are the weaknesses of google search compared to ES?
>
>         Our application works with heavily structured data (large
>         number of small fields) and has lots of faceting and
>         analytics. Is google search any good with it structured
>         documents with many fields? Does it have it advanced faceting?
>         Does it allow fine grained control over indexing?
>
>
>     --
>     You received this message because you are subscribed to a topic in
>     the Google Groups "elasticsearch" group.
>     To unsubscribe from this topic, visit
>     https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
>     To unsubscribe from this group and all its topics, send an email
>     to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="VZCfELGD8noJ">elasticsearc...@googlegroups.com
>     <mailto:<a href="javascript:" target="_blank" gdf-obfuscated-mailto="VZCfELGD8noJ">elasticsearch%2Bunsubscribe@...>.
>     For more options, visit https://groups.google.com/groups/opt_out.
>
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "elasticsearch" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to <a href="javascript:" target="_blank" gdf-obfuscated-mailto="VZCfELGD8noJ">elasticsearc...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

Matt Weber-2
I think you need to focus on your structured data and how you have complete control of how that is indexed/searched with ES.  With the GSA (Google Search Appliance) it is really good at crawling as you might expect.  It can handle structured data but it is quite limited from what I hear.  Based on all the threads and issues I have seen you hoping, you have some very strict requirements on what fields get searched, proximity and boosts between matches across fields, etc.  Meeting these requirements with ES is going to be much easier and quicker than with the GSA.


On Fri, May 24, 2013 at 8:12 AM, AlexR <[hidden email]> wrote:
Thanks Jorg. If it gets that far (hope not) we will do performance trails. I do not think it will have any "cloud" issues because they will run it in the data center...


On Friday, May 24, 2013 5:42:31 AM UTC-4, Jörg Prante wrote:
To add one more important aspect: performance. It should be possible to
order an evaluation license or something for a limited time (1-3
months). Put your data on the product (if possible) and measure the
"documents per second" while indexing and "queries per second" while
searching, for all your workloads. Compare that numbers to ES on your
favorite platform. Add your findings (ease of administration, network
latency, resource sizing challenges etc) Present the result to
management. Management always listens to technical staff when
performance is the issue - it is expensive to fix performance issues later.

Another more important one to kill every cloud-based app scenario is
data privacy but I'm not sure if your data is relevant to this topic.

Jörg

Am 24.05.13 01:10, schrieb Alex Roytman:

> Thanks Tinou,
>
> They have (or would have) google search (I guess appliance) in the
> data center. so i do not think latency will be an issue. it is the
> services it can provide, and integration I need to understand.
> From very briefly going over google documentation my feeling is:
>
> 1.It is not free like ElasticSearch. The licenses may be quite costly
>
> 2.It is optimized for document and site indexing and searching and not
> so much for structured data indexing searching, analytics and faceting
> while elastic allows very easy mixing of both
>
> 3.It is a shrink wrap product which comes with GUI, administration and
> is centered around searching for documents or web pages or emails plus
> some metadata about them not structured data with 100^th of fields
> with need to
>
> 4.It is not an open source or open platform - it is a shrink wrapped
> solution with various connectors to pull data from various data sources.
>
> a.It may be impossible to do something unless google baked it. For
> example showing which fields got hits might not even be possible
>
> b.It will be harder to integrate
>
> 5.Our product will be constrained to what google platform offers and
> we will be bound to versions, licenses and upgrade schedule the IT
> Shop has for the product
>
> 6.All documentation I found is about configuring and managing nothing
> much about developing for it, executing complex custom queries etc.
>
> 7.I checked its user groups – they seems to be dormant unlike elastic
> search. Not sure where one can get any support unless it is paid
>
> 8.Last not the least it will be starting development all over even if
> it is possible to achieve the same with google
>
>
> On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <[hidden email]

> <mailto:[hidden email]>> wrote:
>
>     I haven't used Google search. Briefly played around w/ Amazon
>     CloudSearch. One thing is latency that is probably applicable to
>     any cloud based solution. Within a data center you can update ES
>     and have near realtime updates. With AWS CloudSearch I remember it
>     does things in batches and there's also the latency of a network
>     request to Virginia.
>
>
>     On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:
>
>
>         We have a really good pilot with ES and now infrastructure
>         people pushing google search on us because they have in on
>         their approved stack (and I bet to charge for licenses too)
>
>         What are the weaknesses of google search compared to ES?
>
>         Our application works with heavily structured data (large
>         number of small fields) and has lots of faceting and
>         analytics. Is google search any good with it structured
>         documents with many fields? Does it have it advanced faceting?
>         Does it allow fine grained control over indexing?
>
>
>     --
>     You received this message because you are subscribed to a topic in
>     the Google Groups "elasticsearch" group.
>     To unsubscribe from this topic, visit
>     https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
>     To unsubscribe from this group and all its topics, send an email
>     to elasticsearc...@googlegroups.com
>     <mailto:elasticsearch%[hidden email]>.
>     For more options, visit https://groups.google.com/groups/opt_out.
>
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "elasticsearch" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to elasticsearc...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

AlexR
Precisely my thoughts Matt, thank you! I have already crafted a draft response along these lines just wanted to get some confirmation from the user community that I am not missing existing GSA features and strong points missrepresenting its value in my desire to stick with ES


On Fri, May 24, 2013 at 11:22 AM, Matt Weber <[hidden email]> wrote:
I think you need to focus on your structured data and how you have complete control of how that is indexed/searched with ES.  With the GSA (Google Search Appliance) it is really good at crawling as you might expect.  It can handle structured data but it is quite limited from what I hear.  Based on all the threads and issues I have seen you hoping, you have some very strict requirements on what fields get searched, proximity and boosts between matches across fields, etc.  Meeting these requirements with ES is going to be much easier and quicker than with the GSA.


On Fri, May 24, 2013 at 8:12 AM, AlexR <[hidden email]> wrote:
Thanks Jorg. If it gets that far (hope not) we will do performance trails. I do not think it will have any "cloud" issues because they will run it in the data center...


On Friday, May 24, 2013 5:42:31 AM UTC-4, Jörg Prante wrote:
To add one more important aspect: performance. It should be possible to
order an evaluation license or something for a limited time (1-3
months). Put your data on the product (if possible) and measure the
"documents per second" while indexing and "queries per second" while
searching, for all your workloads. Compare that numbers to ES on your
favorite platform. Add your findings (ease of administration, network
latency, resource sizing challenges etc) Present the result to
management. Management always listens to technical staff when
performance is the issue - it is expensive to fix performance issues later.

Another more important one to kill every cloud-based app scenario is
data privacy but I'm not sure if your data is relevant to this topic.

Jörg

Am 24.05.13 01:10, schrieb Alex Roytman:

> Thanks Tinou,
>
> They have (or would have) google search (I guess appliance) in the
> data center. so i do not think latency will be an issue. it is the
> services it can provide, and integration I need to understand.
> From very briefly going over google documentation my feeling is:
>
> 1.It is not free like ElasticSearch. The licenses may be quite costly
>
> 2.It is optimized for document and site indexing and searching and not
> so much for structured data indexing searching, analytics and faceting
> while elastic allows very easy mixing of both
>
> 3.It is a shrink wrap product which comes with GUI, administration and
> is centered around searching for documents or web pages or emails plus
> some metadata about them not structured data with 100^th of fields
> with need to
>
> 4.It is not an open source or open platform - it is a shrink wrapped
> solution with various connectors to pull data from various data sources.
>
> a.It may be impossible to do something unless google baked it. For
> example showing which fields got hits might not even be possible
>
> b.It will be harder to integrate
>
> 5.Our product will be constrained to what google platform offers and
> we will be bound to versions, licenses and upgrade schedule the IT
> Shop has for the product
>
> 6.All documentation I found is about configuring and managing nothing
> much about developing for it, executing complex custom queries etc.
>
> 7.I checked its user groups – they seems to be dormant unlike elastic
> search. Not sure where one can get any support unless it is paid
>
> 8.Last not the least it will be starting development all over even if
> it is possible to achieve the same with google
>
>
> On Thu, May 23, 2013 at 7:03 PM, Tinou Bao <[hidden email]

> <mailto:[hidden email]>> wrote:
>
>     I haven't used Google search. Briefly played around w/ Amazon
>     CloudSearch. One thing is latency that is probably applicable to
>     any cloud based solution. Within a data center you can update ES
>     and have near realtime updates. With AWS CloudSearch I remember it
>     does things in batches and there's also the latency of a network
>     request to Virginia.
>
>
>     On Thursday, May 23, 2013 12:06:02 PM UTC-7, AlexR wrote:
>
>
>         We have a really good pilot with ES and now infrastructure
>         people pushing google search on us because they have in on
>         their approved stack (and I bet to charge for licenses too)
>
>         What are the weaknesses of google search compared to ES?
>
>         Our application works with heavily structured data (large
>         number of small fields) and has lots of faceting and
>         analytics. Is google search any good with it structured
>         documents with many fields? Does it have it advanced faceting?
>         Does it allow fine grained control over indexing?
>
>
>     --
>     You received this message because you are subscribed to a topic in
>     the Google Groups "elasticsearch" group.
>     To unsubscribe from this topic, visit
>     https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
>     To unsubscribe from this group and all its topics, send an email
>     to elasticsearc...@googlegroups.com
>     <mailto:elasticsearch%[hidden email]>.
>     For more options, visit https://groups.google.com/groups/opt_out.
>
>
>
> --
> You received this message because you are subscribed to the Google
> Groups "elasticsearch" group.
> To unsubscribe from this group and stop receiving emails from it, send
> an email to elasticsearc...@googlegroups.com.
> For more options, visit https://groups.google.com/groups/opt_out.
>
>

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].

For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to a topic in the Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
To unsubscribe from this group and all its topics, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.
 
 
Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

phill
On 5/24/2013 8:29 AM, Alex Roytman wrote:
> Precisely my thoughts Matt, thank you! I have already crafted a draft
> response along these lines just wanted to get some confirmation from
> the user community that I am not missing existing GSA features and
> strong points missrepresenting its value in my desire to stick with ES
>

I also understood that GSA has no ability to deal with access control
which you could provide in an ES product be it in-house or 3rd party.

Another issue is that it brings several parts of the web algorithms to
the enterprise.   At first this sounds like a feature, but it is not
necessarily.

Statistics work great over large populations of search queries when
searching.  An example from today (May 24, 2013) of searching for a word
like "Skagit", or "bridge collapse" (in the news from Washington State
USA) brings up great results on the WWW because of dynamic changes in
searches for those terms etc., but "popularity" and other learning
algorithms can get curious or misleading results in a department or a
division.  This is just one example of where fantastic high-end advanced
algorithms are great for the WWW but don't all apply to enterprise data
sets.

It seems to me that lack of faceting would pretty much be a deal breaker
for me.  In a large enterprise, as a user, MOST of my searches would be
for things in my division or department, I'd really get tired of having
to trick the engine by always remembering to use something from the
path, title, text that IDs my region, division, area of interest etc.

-Paul

--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.


Reply | Threaded
Open this post in threaded view
|

Re: Infrastructure people pushing google search on us:-( any help defending ES would be appreciated!!

AlexR
Excellent observations Paul, thank you. Added to my list.

I believe GSA has pretty good access control but it is based on URL pattern which is pretty useless to us :-)
Yes learning algorithms will be entirely useless to us. we have 


On Fri, May 24, 2013 at 12:31 PM, Paul Hill <[hidden email]> wrote:
On 5/24/2013 8:29 AM, Alex Roytman wrote:
Precisely my thoughts Matt, thank you! I have already crafted a draft response along these lines just wanted to get some confirmation from the user community that I am not missing existing GSA features and strong points missrepresenting its value in my desire to stick with ES


I also understood that GSA has no ability to deal with access control which you could provide in an ES product be it in-house or 3rd party.

Another issue is that it brings several parts of the web algorithms to the enterprise.   At first this sounds like a feature, but it is not necessarily.

Statistics work great over large populations of search queries when searching.  An example from today (May 24, 2013) of searching for a word like "Skagit", or "bridge collapse" (in the news from Washington State USA) brings up great results on the WWW because of dynamic changes in searches for those terms etc., but "popularity" and other learning algorithms can get curious or misleading results in a department or a division.  This is just one example of where fantastic high-end advanced algorithms are great for the WWW but don't all apply to enterprise data sets.

It seems to me that lack of faceting would pretty much be a deal breaker for me.  In a large enterprise, as a user, MOST of my searches would be for things in my division or department, I'd really get tired of having to trick the engine by always remembering to use something from the path, title, text that IDs my region, division, area of interest etc.

-Paul


--
You received this message because you are subscribed to a topic in the Google Groups "elasticsearch" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/elasticsearch/cbP2lp5UtZI/unsubscribe?hl=en-US.
To unsubscribe from this group and all its topics, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.



--
You received this message because you are subscribed to the Google Groups "elasticsearch" group.
To unsubscribe from this group and stop receiving emails from it, send an email to [hidden email].
For more options, visit https://groups.google.com/groups/opt_out.